Best Play A New Method For Discovering The Strongest Move Alexander Shashin 2
Best Play A New Method For Discovering The Strongest Move Alexander Shashin 2
Best Play A New Method For Discovering The Strongest Move Alexander Shashin 2
PART II. A Universal Method for Discovering the Strongest Move, in Practice
BEST PLAY:
© 2013 Alexander Shashin
All rights reserved. No part of this book may be reproduced
or transmitted in any form by any means, electronic or
mechanical, including photocopying, recording, or by an
information storage and retrieval system, without written
permission from the Publisher.
Publisher: Mongoose Press
1005 Boylston Street, Suite 324
Newton Highlands, MA 02461
info@mongoosepress.com
www.MongoosePress.com
978-1-936277-46-9 paperback
978-1-936277-47-6 hardcover
Library of Congress Control Number: 2013935088
Distributed to the trade by National Book Network
custserv@nbnbooks.com, 800-462-6420
For all other sales inquiries please contact the publisher.
Editor: Jorge Amador
Layout: Andrey Elkov
Cover Design: Al Dianov
Printed in China
First English edition
0987654321
Foreword
My first acquaintance with A.A. Shashin took place in the spring of 2003, at a very difficult point in my career. The
year 2002 had been one of the most unfortunate in my entire 15 years of professional activity. My rating had dropped
almost 80 points, to a level (humbling for me) of 2679; my world ranking fell to twentysomething, and I hardly knew
what to do next. Fate had it that I should then meet this amazing person, coach, and physicist by training (and perhaps
by calling): San Sanych Shashin, as I freely started to call him.
It is difficult to overestimate the amount of support he gave me. Our many hours of kitchen-table conversations brought
me back to life as both a chessplayer and a human being.
Thanks to his patience and his extraordinary level of native intelligence, San Sanych succeeded not only in renewing my
appetite for chess, but also in showing me those edges of life which until then – as a result of my age and the
pecularities of my character – I had steadfastly ignored.
In a purely chess sense, the question logically arises: could a Soviet master with thirty years’ experience as a trainer, and
master of his own attitude in the search for a decision-making algorithm, talk about chess as an equal with someone
who until recently had ranked in the Top Ten in the world, and still teach him something?
I reply: both yes and no. “No,” because it is difficult to remake an already fully formed player, 26 years old and
successful as a result of his own talent and his chess understanding. But “yes,” because any crisis is an opportunity for
growth – and I, discovering the conclusions of S.S.’s theory at precisely that moment, significantly enriched and
broadened my horizons: while selecting a move, I often successfully employed the ideas he presented to me. The fact
that the period from July 2003 to July 2004 was the most successful of my entire career, I owe in great part to our
kitchen-table discussions. I returned triumphantly to the Top Five, along the way winning practically every tournament
I participated in.
At a grandmaster conference on the “Morozevich problem,” assembled spontaneously during the 2004 Russian team
championship in the bar of a hotel in Dagomys under Alexander Khalifman’s chairmanship (who, at that tournament,
declared in an interview that, in terms of playing strength, Morozevich was now “number one” in the world), various
guesses of the most varied degrees of relation to reality were proffered for this. But, as you know, stern realities are
always more prosaic than flights of fancy, and the role of the humble master Shashin was noted by no one…
Despite nearly a half-century of work in the field of chess, San Sanych remains unknown to most chess fans, not only
around the world, but even in Russia. Remarkably modest by nature, he never sought fame or any increase in his sphere
of influence, and he practically never gave interviews. It is not surprising that, for most people who had at least heard of
him, he is viewed as a sort of hermit who became somewhat known to the public only thanks to the publication of
particular articles on the website “e3-e5” and to previews of this book on the “bs-chess” site. Far from a scientist-
wizard who sits at home and feverishly concocts hard-to-understand theory, San Sanych is actually a man with well-
defined views on life, able to take a principled stand on many issues and to defend it. Few know that, in the 1970s, it
was precisely Shashin who did not fear to give public support to the blackballed Victor Korchnoi, although even then
his favorite chessplayer was Karpov. It was simply that the inhuman persecution of the former was stronger than
Shashin’s mere chess sympathies. This book sheds light on the entire creative path of Shashin as a chessplayer, as a
trainer, as a physicist and theoretician, and as a very brave man, trying to put together his knowledge into a complete
and dynamic system, which (like any conceivable theory) cannot explain the unfathomable and understand the core of
the phenomenon of chess. The attempt itself, the acceptance of the challenge, is worthy of admiration. Fewer and fewer
are the people who are ready to write and to discuss the “eternal” themes, and fewer still who even think about such
questions.
It is no doubt impossible to write in a simple manner about the model of chess as the physicist Shashin sees it. However,
the difficulty of taking up and studying it should in no way scare off the reader! In this case, the author’s idea is not to
squander his effort in trying to set out, bit by bit, his own knowledge and in acceptable form to bring it to us, but to
make us think of the depth and inexhaustibility of chess, of its historical roots; and most of all, that even in our super-
advanced computer age the chief secret of chess as it is played – the search for an algorithm for finding the best move –
is still unresolved. Many people forget this. And along with that, it is still beyond the scope of the chess-reading
public’s interest. Moving from the 6-piece tablebase to the 7-piece one in parallel with the further endless plunge into
the opening jungle – practically speaking, that’s all that chessplayers concern themselves with nowadays. I especially
wish to recommend this book to young professional chessplayers just starting out: for you it is also necessary to obtain
a higher education, as without both a serious chess and general education, individual victories will not make you a true
sportsman or human being!
I want to thank A.A. Shashin once more, from the bottom of my heart, for writing this outstanding work, and to remind
us that this book is only one stretch of that endless road under the signpost of chess… New heights to You, Master!
Alexander Morozevich, International Grandmaster
Introduction
Dear reader, in your hands you are holding a somewhat unusual chess manual. Let me be more accurate: this manual is
completely unusual. How, or why?
Because, by studying it, you will learn an original protocol for identifying the strongest chess move in any position, one
which has nothing in common with traditional techniques. I have called this method “universal,” and I have no doubt
but that it will help us to find the strongest move in all possible chess positions, without exception – be they positions
with all 32 pieces on the board or positions featuring the barest minimum of pieces.
This book is split into two parts; the first is the elements. The most important section of the first part of this book is
Chapter 6, where we find all the most important components of the universal method for searching for the strongest
move.
The second part of this book develops naturally from the first. There, you will find examples of varying degrees of
complexity. In my view, the first two chapters in Part II will be within the limits of the powers of a strong Russian
second-category player (in U.S. terms, a Class B player). Chapters 3, 4, and 5 are more complex, but in the end, “if
there’s no pain, there’s no gain!” You can do it! You will get through these chapters also.
The first 125 examples were annotated with the help of chess engines. These were various permutations of Fritz
(particularly Fritz 11) and the program Rybka 3 (32-bit). My hope is that this will keep the number of gross analytical
blunders to a minimum.
For readers who lack access to a computer: Rest assured, in the larger scheme of things, we don’t need a computer! I
repeat: It’s not necessary!
Why?
The reason is simple. Mine is a method for searching for the best move by a human, not by any of our silicon friends.
Let me add to that: the universal method for identifying the best chess move is, to all intents and purposes, a three-part
method derived from the ideas of three extremely gifted players, all of whom lived in the pre-computer era. They were
Mikhail Tal, José Raúl Capablanca, and Tigran Petrosian.
This method is the first of its kind. The very best move is a child of the method. Tal, Capablanca, and Petrosian are our
teachers. Fritz and any flavor of Rybka are merely our assistants.
I cannot conclude this introduction without saying, “Thank you.” I am grateful to many people (both chessplayers and
non-chessplayers) for their help.
Three especially: Alexander Kentler, Leonid Yudasin, and Vladimir Bazhenov. The first – for the chance to publish, in
the magazine Shakhmatny Peterburg, a series of articles concerning this method. For the second – we are speaking of
international grandmaster Leonid Yudasin, author of the fundamental investigation The Thousand-Year Myth of Chess
– for the opportunity to put in his book an article on the problems of the foundations of the game of chess that was very
important to me.
I want to single out Vladimir Bazhenov for special mention. Most of all, for engaging me in a deep discussion over the
number of parameters in a chess position. The result was that I had to increase the number of parameters from four to
five.
Really, the game of chess is practically inexhaustible. Chess is limitless…
A.Shashin
St. Petersburg, January 2013
Part I
Chapter 1
An Overview of Part I
Our ultimate goal in chess, which we will resolutely pursue throughout Part I, is a universal method for discovering the
strongest chess move. More than that, a method that works in all possible chess positions, without exception. In all of
them!
Without a doubt, we will achieve this goal. This will happen at the very end of the first part of the book, in its sixth
chapter. Chapter 6 is the most important one – why?
Because in that chapter we gather all the elements in one place, we review, summarize, and sort out all the chess wisdom
in the preceding chapters. Everything comes together in Chapter 6: without this, we do not have a universal method.
The sum total of all our work is the chart (given at the end of this chapter and also in Chapter 6) showing how the
algorithms for discovering the strongest chess move flow, or drift, into and from each other. This “Algorithm Drift
Chart” is the meat of the universal method. It is this model which is, in fact, the purpose of our chess efforts.
Our task now – a task which we will work out over the course of Part I – is to “decode” the model and tease out its rich,
inner substance.
You may have already come across the relevant page and had your first exposure to the Algorithm Drift Chart... What
does it mean? What do you see there?
There, you will see three “zones,” located along the “t” axis. Then too, there are the five parameters – from “m” to
∆(move). Two original “baskets” of information...
We start with a “basket” full of parameters: there are five of them. These are: “m,” “t,” a “hidden” parameter – the third
in order – which we shall be naming shortly, and parameters ∆k and ∆(move). What is all of this?
This “Algorithm Drift Chart” is the meat of the universal method. It is this model which is, in fact, the
purpose of our chess efforts.
I reply: these are the parameters of any given chess position. Five numbers. And these five numbers comprise a value.
Every position has its own value, its own stamp consisting of these factors...
In this model, what do these five parameters mean?
Briefly, this is what they signify:
1) The “m” parameter is the material factor of the chess position;
2) The “t” parameter is the factor of time (piece mobility) in chess;
3) The third parameter is the factor of safety in a chess position;
4) Parameter ∆k stands for the first of two space factors of a chess position; and
5) Parameter ∆(move) is the second space factor.
That’s enough for now. Suffice it for us to recognize just two simple ideas: 1) there are five parameters in every position
(I repeat: in every single position, since chess positions exist in time and in space); and 2) five parameters → five
numbers → one value → one “zone” or another (see the drift chart).
Furthermore, no longer will a chess position be left to its fate. There shall be no “orphan” positions. All chess positions,
without exception, will find their proper “zone.”
The next step is the second “basket” – a bottomless basket filled to the top with trillions upon trillions of chess positions –
a countless number... One thing which all of these positions in this bottomless basket, split into three “zones,” have in
common is that the search for the strongest move in all of these “zones” follows a specific algorithm.
Positions in the “Tal Zone” correspond to the Tal Algorithm (the algorithm for attacking material chess targets). We’ll
encounter the Tal Algorithm mainly in Chapter 2. For us, “Tal” means the search for the strongest chess move when
attacking.
“Tal” means the search for the strongest chess move when attacking.
“Capablanca” means searching for the strongest move in strategic play.
“Petrosian” covers the search for the strongest move in defense.
Together, the three algorithms (T, C, and P) encompass the whole, undivided spectrum of all chess attacks
and defenses.
Next, the “Capablanca Safety Zone” corresponds to the Capablanca Algorithm (the algorithm for discovering the
strongest strategic move). We will learn about the Capablanca Algorithm in Chapter 3. For us, “Capablanca” will mean
searching for the strongest move in strategic play.
Finally, the “Petrosian Zone” (Chapter 4) covers the search for the strongest move in defense. Together, the three
algorithms (T, C, and P) encompass the whole, undivided spectrum of all chess attacks and defenses.
These are fundamental algorithms. The entire wisdom of the game rests upon them. They are what give us hope, drawing
us closer to our dream – the dream of identifying the strongest move in any position.
Moving on, we get to Chapter Five. What’s there?
There, we find the “mixed” algorithms – the TC, CP, and TCP algorithms for determining the strongest chess move. The
TC and CP algorithms occupy the gray areas between neighboring algorithms, while the TCP Algorithm describes
complex (and extremely complex) positions. Though not fundamental, these three algorithms are still of exceptional
interest, as we’ll see!
The five parameters of the chess position, the fundamental and non-fundamental algorithms, the Algorithm Drift Chart
for searching for the strongest chess move – all this is our chess environment. We must live in this environment and
make it our home.
An Eastern saying goes: “A journey of a thousand miles begins with a single step.” We will be taking that step very soon
– the first step on the road to our dream, along the endless road of chess...
Leading us will be the strongest of the strong among chessplayers. Our gratitude to them is infinite. If we wish to be their
worthy pupils, then we must serve the objective of chess truth unreservedly – truth in chess lies in the strongest move!
Are you ready?
The five factors and the five parameters of any chess position:
1) the material factor of a chess position: the “m” parameter;
2) the factor of chess time: the “t” parameter;
3) the factor of safety in a chess position, and its parameter;
4) the ∆k parameter: the factor of compactness in a chess position;
5) the ∆(move) parameter: the factor of increased space (expansion).
If this chart doesn’t make sense right now, that’s OK. You will find many references to it in the coming pages, so we
include it here so that you can consult it and start figuring things out.
Chapter 2
No. 1: Tal – NN
Simultaneous exhibition, Stuttgart 1958
Position after 13...Ne5
In this position, it is White to move, and we are playing White. What do we have?
We have an open attack! And precisely:
14.g5
With this move, White attacks two material chess targets at the same time – one on f6, and the other on h6. White
threatens to win material (threatening 15.gxf6 and 15.gxh6).
An “open attack” is always a one-move threat. In terms of time, this is the shortest route to success.
14...hxg5 15.hxg5
The knight on f6 is hanging – another open attack! In addition to this, White offers his opponent the opportunity to trade
rooks.
To trade or not to trade?
15...Rxh1
I won’t criticize Black’s play. Our theme is not the algorithm of defense, but the algorithm for the attack on material
chess targets. Our theme right now is “Tal,” not “Petrosian.”
16.gxf6
16...Rxd1+
Check! Check is always an open attack on the king, a one-move threat. Its distinctive feature is that it is an assault on a
piece that possesses infinite chess value.
17.Nxd1
The white queen joins in the assault – another open attack. The target is the piece on a5. And let’s not forget the target
on g7!
17...Qxd2
18.fxg7
As we can see, White has delayed “until further notice” the act of removing the opposing queen. He has come up with a
stronger threat – the terminal prospect of 19.g8Q#. Checkmate is just a radical type of check. Mate is always a one-
move chess action. With mate, the enemy king is guaranteed to be destroyed in one move. The threat of immediate mate
is a two-move action and a type of attack on the king that we will label a direct attack on it.
Note that, in the final position, the chess value of the empty square g8 is just as great as the chess value of the occupied
square e8. As the black king (whose value is infinite) sits on e8, this means that these values are limitless.
1-0
Postscript
An important question for us – what was the basic reason for Black’s crushing defeat?
Short answer: White’s overwhelming advantage in the number of forces participating in the attack!
Now let’s answer the question in more detail.
In his offensive, White threw the g-pawn into the battle. One move later, the h-pawn was activated as well. The file
came open, and the black rook, standing on h8, went to h1 (White sacrificed the rook). Then, after the rook trade on d1,
White’s queen entered the fray from the d2 square.
Finally, a queen sacrifice, and... the g7-pawn was unstoppable. Victory!
We readily see that, though brief, this was a bloody chess battle. In the fight there perished: White’s g-pawn, two rooks,
and queen; and Black’s h-pawn, knight, rook, and dark-squared bishop. They gave up their existence on the squares g5,
h1, d1, d2, f6, and g7.
Question: for what purpose did Tal arrange such a battle?
The answer is obvious: for the sake of the target on e8! The opposing king is always Target Number One. The king is a
target of unlimited chess value...
The squares e8, g5, h1, d1, d2, f6, and g7 – these are inner and outer squares of the rectangle d1-d8-h8-h1. The main
events take place within this rectangle and on its perimeter. In other words, the rectangle d1-d8-h8-h1 is in fact, the area
of attack – the sector where Tal tore apart his hapless opponent.
In this attacking sector, Tal had an extra queen (see the starting diagram). More precisely, White’s queen and rook are
opposed by king, knight, and pawn. This means that White’s superiority was overwhelming.
Another thing: keeping in mind the king’s unlimited value, note the squares within the rectangle d1-d8-h8-h1 that are
immediately adjacent to the king. These are the squares of the rectangle d7-d8-f8-f7, and it’s clear to us that these
squares are safely protected. In other words, Black’s king is safely “cocooned” – its faithful co-workers occupy the
squares d7, e7, and f7. However, it is equally true that, in the rectangle d1-d8-f8-f1, White is up a queen and a rook...
This position came up after 1.c4 c5 2.Nc3 Nc6 3.Nf3 Nf6 4.e3 e6 5.d4 d5 6.cxd5 Nxd5 7.Bc4 Nb6 8.Bb5 a6 9.Bxc6+
bxc6 10.0-0 Bb7 11.Ne4 Nd7 12.Qc2 Qb6 13.Ne5 cxd4 14.Nxd7 Kxd7 15.exd4 Ke8 16.Be3 Qc7.
We see that Black has already lost the right to castle after 14...Kxd7. That can’t be good...
In the diagram position, it is White to move – and we’re playing White. What have we got?
We clearly have even material, and an obvious advantage in the time factor.
Question: Why does White possess this obvious advantage in time?
Answer: Because the white king, queen, bishop, and knight are already developed: these pieces are already in action.
They are four in all (we will consider castling as only one move; the rook on f1 doesn’t count). Therefore, we add 4 to
our right to make the next move (1)!
Meanwhile, what does Black have?
Black only has a bishop and the queen in the battle. Do you think that’s too little?
Simple arithmetic: 4 + 1 – 2 = 3. White has three extra tempi. That’s a solid advantage, which impels us to active play.
However, we must acknowledge that our arithmetic is quite crude: it doesn’t always hit the mark. Our calculation is but
a rough approximation of the truth. Tempi are “rough and concrete” (to quote the poet Mayakovsky). Truth lies in the
arithmetic of the mobility of the chess pieces...
So, how do things stack up in that regard?
We have 44/35 in favor of White, where 44 and 35 are the sums of the mobility of all the white pieces and all of the
black pieces (see the postscript to this game). This means that the factor of time in the current position does indeed
impel us to take active measures!
In addition, the safety factor also pushes us into action (for this, also see the postscript). It is more than obvious that
Black has a “bad” king – he literally “attracts” enemy pieces to himself.
Tal delays for not a single second!
17.d5
17...exd5 18.Rfe1
This move is not only fearless, but – I say – also correct. With this move, Tal brings the formerly inactive rook into the
field of battle. The second and third points of the attacking algorithm (for attacking material chess targets, the Tal
Algorithm) are at work, and we have harmony!
18...Kd8
Fritz considers this best. 18...dxe4 would be utterly bad: after 19.Qxe4+ Be7 20.Bc5 Black is helpless, as he has lost the
right to castle.
19.Qb3
Threatening 20.Bb6 winning the queen. The queen is Target No. 2 – Target No. 1 being the king. The queen is a piece
of great but finite chess value, while the king has infinite value.
19...c5
A tactical error that merely hastens the inevitable. 19...Kc8 would have dragged things out a bit. However, here too the
computer coolly predicts victory. A sample variation: 20.Nc3 Qd8 21.Rac1 Kb8 22.Bb6 Qd7 23.Nxd5 – and I’ll let
you figure out the rest!
White’s attack was absolutely correct, not simply because he had a clear edge in the factor of chess time, but precisely
because (apologies for the repetition): 44 > 35, where 44 = 1 + 12 + 4 + 4 + 5 + 7 + 11, and 35 = 3 + 12 + 4 + 1 + 1 + 5
+ 9.
Here the numbers 1, 12, 4... represent the number of squares to which the white king, the white queen, the rook standing
on a1, and so forth, can move (their mobility), ending with the number of possible moves by the white pawns.
Meanwhile, the numbers 3, 12, 4... represent the mobility of the black king, black queen, a8-rook, etc.
The position after 16...Qc7 yields a strong advantage for White.
The chess value scale: king, queen, rook, minor piece (bishop or knight), pawn.
When attacking our opponent, we strive first to attack the king, then the queen, next the rook, and so on. In
other words, we attack our opponent’s pieces according to their chess value.
Why?
First of all, because the “t” parameter in this position is greater than 1 (t = 44/35, or ~1.26), meaning that White’s pieces
enjoy greater mobility than do Black’s. And secondly, because the black king is in danger. Indeed, the king on e8 is
hardly securely placed; and besides, as is not difficult to see, within the d1-d8-f8-f1 rectangle, White has an extra rook
and knight.
12.Nf5
Sacrifice! Tal sacrifices a piece to open the e-file, in order to bring the rook to the e-file with tempo.
Question: Is the sacrifice correct?
I’m not going to answer that one. I am decidedly silent. I will say only that Tal’s fearlessness sometimes bordered on
recklessness.
Your task is to put this position under computer analysis!
12...exf5 13.Re1+
13...Be6 14.Qd6
The queen – the most powerful piece on the board – is thrust into the attack.
14...a6 15.Bd2
By retreating, we attack! The queen is always Target No. 2. The queen has a value smaller only than that of the king.
Target No. 3 – that’s the rook. There follow the bishop, the knight, and then the tiniest chessman, the pawn.
15...Qxc2 16.Bb4
There is no satisfactory defense against the check on f8. Uhlmann resigned after
33.R1h6
Tal throws his last reserve into the attacking zone. He threatens the crushing 34.Rf6+ gxf6 35.Bh6+, etc. Can Black
hold?
Master Koblents couldn’t find a way. He replied
33...d3
34.bxc3
34...d2+ 35.Kd1
Of course!
35...Qxc6
Too late!
36.Rf6+ Rf7
37.Qxg7+
1-0
Chess is a most complex, practically inexhaustible game. Chess science cannot explain everything by
breaking it down to little pieces. In chess, there are many exceptions to the rules. It’s unpredictable – and
fortunately so!
Chess is a most complex, practically inexhaustible game. Chess science cannot explain everything by breaking it down
to little pieces. In chess, there are many exceptions to the rules. It’s unpredictable – and fortunately so!
We do have a compass: the Tal Algorithm. This algorithm will help us chart a true course in our future chess ocean.
One more thing: Remember this, reader, for your long chess life: Chess loves a fearless hero!
A second point of the Tal Algorithm is that the pieces are always ideally placed on squares which coordinate with later
open and direct attacks on the opponent’s material targets. Most often, a rook is thrown into the attacking zone, much
as in Tal – Koblents. However, sometimes you can throw the rook... without throwing it!
18.Kh1
18...Bd4
If 18...fxg5, then, according to Tal’s analysis, White wins with 19.Rf3+ intending 20.Bxe5. Let me add a personal note:
still stronger is the immediate 19.Bxe5 (pointed out by Rybka).
19.f4
Now the rook, though remaining on f1, becomes a full participant in the attack on the king.
19...Qb6 20.Bc3
Find the win after 21...fxg5. And you are not allowed to use a computer!
22.fxe5
The beginning of the end.
22...Nxe5 23.Ne4 h5
24.Nxf6+
Check – that is, an open attack on the king. Van der Wiel resigned, as he cannot avoid mate after 24...gxf6 25.Qxe5.
Our ideal goal in chess is error-free play in all possible positions, without exception. An unattainable goal, but the desire
to achieve it is eternal.
Our ultimate chess objective is to have a universal method for finding the best move in all possible positions, without
exception. The goal is a threefold one, and its facets are the fundamental algorithms of Tal, Capablanca, and Petrosian.
Our chess goal so far is the Tal Algorithm, which we have followed, are following, and will continue to follow in the
immediate future. This is why, first of all, in this chapter we are interested in the method (the algorithm) itself. And
only secondarily will we be interested in finding the strongest move.
On the other hand, it’s crystal clear that the latter goal (finding the strongest move) is nearly impossible to attain without
the former (understanding the algorithm for accomplishing
that).
Our immediate chess objective – that is, our goal “here and now” – is to break down another Tal game. This will be the
game against Tolush, and in our analysis of this game Garry Kasparov himself will help us. The former world
champion has analyzed this rather complicated game in great detail (see My Great Predecessors, Part II). This game
was a mind-bender, an irrational event...
Ready?
15.Bb5
The text move is an open attack on the queen, which White combines well with the injection of a recently inactive piece
into the conflict zone, and sacrificing it. Impressive, no?
Nevertheless, Kasparov disagrees with Tal, and thinks that Tal erred. He writes: “Intuitively feeling that the position
demands a sacrifice, Tal immediately sacrifices – at just the right moment, but... the wrong piece and on the wrong
square! Soon afterwards they found 15.Nxe6!! fxe6 16.Nd6+ Bxd6 17.Qxd6 Rf8+ 18.Kg3 with a very strong attack...”
I hope that you, dear reader, will not have forgotten the value scale for attacking material chess targets: king, queen,
rook... The king is Target No. 1, the queen No. 2, etc.
That means that, when studying the position, we cannot ignore 15.Nxe6. Why?
Because after 15.Nxe6 White threatens the fatal 16.Nc7#.
The question is: which is worth more – a direct (2-move) attack on the king, or an open (immediate) attack on Target
No. 2?
Don’t answer yet. It’s not that easy!
15...axb5 16.Nxb5
Indeed it is not so easy, since after the text move he also threatens a killer knight move to c7!
With this move, Tal throws one knight into the conflict zone, formally sacrifices the other one (on e4), and informally
regains a portion of the sacrificed material.
16...f6 17.exf6
Kasparov: “Also giving up the knight at e4. The standard evaluation of the position has to give way to intuition and
calculation. There is no mate, and it would seem that, based on the strict canons of higher chess truth, White must
sooner or later lose. But what is the practical significance of this ʻhigher chess truth,’ if the variations confirming it are
found only months or years later?!”
Tal’s move is not just a sacrifice. He takes a pawn, as well as openly (or directly?) attacking. Or perhaps White merely
threatens an innocent trade on g7? Is Tal playing strategically – yes or no?
Tolush took that pawn on f6. There followed
17...gxf6
Kasparov dedicated over two pages in My Great Predecessors to analyzing the position after 17.exf6. Chess is truly
inexhaustible! His verdict: Black’s last move was a mistake.
The “greedy” 17...Qxe4 looks stronger. And, strangely enough, after 18.fxg7 Bc5+ 19.Kg3 (see diagram), Black
manages to hang on:
18.Re1
With this, Tal develops his formerly inactive rook – the second point of the algorithm. And let’s not forget the piece on
g5, which is to be sacrificed!
18...Ra6 19.Bxf6
19...Nxf6 20.Nxf6+
An open attack on the king + the win of material (was it a trade, after all?). The first and second elements of the
algorithm intertwine. So, does this make it a trade? Yes or no?
20...Kf7
21.Rf3
We’ve already seen this. Which is to say: White uses a standard method for adding to the attacking zone a piece that is
currently doing very little. This rook move may be made (and Tal did make it) instantaneously, automatically – and
why not: the threat of discovered check is extremely uncomfortable for Black! I add: I chose this move instantly, also
without thinking to analyze the position; so did Kasparov...
However, soon after publication of the second title in the My Great Predecessors series, it was shown that the rook
move is a mistake. Better is 21.Ne4, as pointed out by Sorokhtin – an expert-level player! After 21...Be7 22.Rf3+ Kg6
23.Qf4 or 21...Nd7 22.Qg5 Be7 23.Qh5+ Kg7 24.Rg3+ Kf8 25.Rf3+ Bf6 26.Qg5, etc., White should win.
Is it possible that these two great players missed this possibility?
Of course, I haven’t even a single thought of doing anything against the legacies of Tal or Kasparov. They are immortal,
and they need no bootlicking. They probably are not gods, but they are demigods, at least... And still, very great
chessplayers are known to make mistakes!
I don’t work to idealize these great, great players. I work for you, dear reader, and I do have a dream. I dream that
someday, some unsung NN – today, a chessplayer of no very great reputation – will read these lines. He will read them
and... will become, eventually, the World Chess Champion!
Sorokhtin’s move, 21.Ne4, probably wins. Isn’t it because it is multifaceted (checks are threatened by the queen and the
rook and the knight)? And one more thing: this move cannot be pigeonholed into one particular category, as it contains
elements of both direct attack and optimal placement. On the other hand, we are already used to the fact that in Tal’s
games, he commingles everything. All four elements are mixed up with one another...
Question: does Tal always play “according to ʻTal’”? Or: does the Tal Algorithm have a chance to become our universal
algorithm? Yes or no?
21...Qh4+ 22.Kf1
22...e5
23.Qd5+
The queen attacks the king – the most powerful of all assaults in chess.
23...Be6
24.Nd7+
The next open attack on the king. Moreover, I should add, the climax of a grand struggle!
24...Kg6
As we say in Russia, “Wrong steppe!” Kasparov: “After 24...Ke7 the wretched king, supported by its subjects, could
have hoped to survive the storm. It looks like White has nothing better than 25.Qc5+ Ke8 26.Rxf8+ (26.Qc8+ Ke7;
although I add: after 27.Qxb8 Bg7 28.Qxb7 Qc4+ 29.Kg1 Qc8 30.Qxc8 Rxc8 31.Nxe5 Rc5 32.Rf7+ Bxf7 the
machine says White has the advantage – A.Sh.) 26...Rxf8+ 27.Qxf8+ Kxd7 28.Rd1+ Kc6 29.Qd6+ Kxb5 30.Rb1+ Ka4
31.Ra1+ Kb5 32.Rb1+ with perpetual check.”
In other words, according to Kasparov’s analysis, Tolush’s last move was a mistake. More accurately: the last mistake
Black made in this game!
25.Nxe5+
25...Kg7
26.Rg3+
An open attack.
26...Qxg3
27.Qxb7+
27...Nd7 28.hxg3
In the aftermath of a chess battle, the field is littered with corpses and mangled equipment.
The elements of the Tal Algorithm:
1) open and direct attacks;
2) optimal piece placement;
3) sacrifice;
4) winning material.
The first two elements define the dynamic and the potential (strategic) components of the attack.
The last two are an unequal trade with its mirror image – an exchange.
The game concluded with:
14.dxe6
Sacrifice – the third point of the algorithm for attacking material targets.
I add: Kasparov is in complete solidarity with Tal. He considers the text to be strongest and gives it an exclamation
point.
14...bxc3 15.Nd4
Kasparov: “Apparently the strongest continuation, although White could also have considered 15.bxc3!?...” In other
words, after 14...bxc3 White has a choice, which means that the second (15.Qd4) and third (15.bxc3) elements of the
algorithm are at odds with each other.
15...Rg8
Kasparov doesn’t hide his enthusiasm for this move. In his view, 15...fxe6? and 15...Qc6? are both bad, although
15...Bd5 is entirely possible. After each of these moves, there are variations piled on variations...
Here’s one of them, a truly fantastic one: 15...fxe6? 16.Nxe6 Qd7 17.Bh5+ Ke7 18.Qg4! Bd5 19.Rfe1! Bxe6 20.Rad1
Qc6:
16.Qa4+
An open attack on the most important target in chess. Remember the scale of targets: king, queen, rook, bishop, knight,
pawn... The king is Target No. 1; Tal resolutely attacks it.
16...Kd8 17.g3
And this move is something indescribable! Hard to believe, but Tal is... defending.
Why?
Because, sadly, the superbly valiant 17.Rfd1 fails to 17...Rxg2+ 18.Kf1 Rxf2+, etc.
17...Bd5
Kasparov neither criticizes nor praises this move. Other moves are possible: 17...Nc6?, 17...Bc5!?, 17...cxb2, and
17...Rg5. The first of these four loses, while the rest lead mainly to unclear play. And everywhere variations, sub-
variations, sub-sub-variations...
18.Rfd1
Adding the rook into the attacking zone, yet a small inaccuracy. Stronger is 18.Rad1, when after 18...cxb2 the rook
doesn’t hang, and White gets a very strong move: 19.Bf3! (Kasparov).
My promised computer correction of the former world champion’s computer annotations: 19.exf7 was stronger than
moving the bishop to f3; if 19...Qxf7, then 20.Bxc4 Rg5 21.Bxd5 wins (Fritz 11).
18...Kc8
“A serious mistake” (Kasparov). In his opinion, ...18...cxb2 19.Rab1 c3 is far stronger. Dynamic equality!
19.bxc3
Kasparov considers this unsatisfactory and gives it a question mark. White wins not by taking the pawn, but with the
check 19.Qe8+! and then 19...Kb7 20.bxc3 Nc6 21.Rab1+ Ka7 22.exf7 Rg5! 23.Nf3! Ne7 24.Nxg5 fxg5 25.Bf3!+-.
“However,” writes Kasparov, “it is hard to criticize Tal: this win is not obvious at all, and at any rate it was not found
by three generations of analysts, until computers became involved ...”
Let me add that a wide range of illustrious players – including Euwe, Hübner, and Timman among others – have
analyzed this game or this position at one time or another.
19...Bc5
20.e7
Outstanding! White’s thrust was not simply a pawn sacrifice to accelerate the attack. It was also a direct attack – the
empty square e8 is attacked (21.e8Q+ is threatened). With this move, Tal harmoniously combines the first and third
elements of his algorithm.
20...Nc6 21.Bg4+
21...Kb7 22.Nb5
But this rambunctious invasion he frowns on...
In Kasparov’s view, 22.Nxc6 is bad. 22.Nf5!? – openly attacking Black’s light-squared bishop – looks much better. But
22.Nb3!? is even stronger (all marks are Kasparov’s). Putting the knight on b3 is a double attack (an open attack on
Black’s dark-squared bishop, plus an open attack on the light-squared bishop). In addition, White would be sacrificing
this knight. And don’t forget the g4-bishop, which is still standing!
22...Qe5
23.Re1
An open attack on the queen. Target No. 1 is out of reach for the time being.
23...Be4
24.Rab1
24...Rxg4
According to Koblents (Tal’s trainer and a longtime friend), on 24...Bxb1 Tal had a miracle move up his sleeve:
25.Rxb1!. Kasparov quietly approves of this intuitive inspiration and projects a win for White after 25.Rxb1. If only, if
only...
25.Rxe4
25...Qxe4
25...Rxe4?! looks inferior. After 26.Nd4+ Nb4 27.Qd7+, White has the upper hand (according to Kasparov).
26.Nd6+
“With a time scramble approaching, Tal finds literally the last attacking resource, and the heroic Keller, who has made
so many strong moves, is unable to withstand the extreme tension” (Kasparov).
28...Re5
29.Rb7+
An open attack on the king + a sacrifice! This is the first of a winning series of checks.
29...Kxb7 30.Qd7+ Kb8 31.e8Q+ Rxe8 32.Qxe8+ Kb7 33.Qd7+ Kb8 34.Qxc6, and Keller gave up. 1-0
Let’s proceed to the post-mortem, including a comparative parametric analysis. We will examine the positions after both
13...Qc7 and 28...Re5 (see the first and the last diagrams above; in the diagram after 28.Qd1, don’t forget to move the
black rook from e4 to e5).
We know the parameters of the position after 13...Qc7: a deficit in the material factor, approximate equality in the time
factor (t = ~1.03), and a substantial advantage in the safety factor. Or, briefly: m < 1, t ≈ 1.03, “+”.
Here, m < 1 indicates a deficit in the material (or first) factor of the position. (And m = 1 and m > 1 mean, respectively,
positions with material equality and positions with material superiority.)
The symbol “+”, it seems to me, speaks for itself and needs no further comment. Its colleagues are the symbols “=” and
“-”.
And now for the parameters of the position after 28...Re5: m < 1, t = 36/38 = ~0.95, “+”. Work them out!
If my basic arithmetic doesn’t deceive me, then, roughly speaking, we have the same “diagnosis.” In both positions, we
have the same value of “m,” a “t” parameter close to 1, and a solid “+” according to the third factor. In other words, we
have a serious basis for rewriting our previous diagnosis. “Rewrite” from the position after 13...Qc7 to the position
after 28...Re5. For us it is transitional (or mixed), since back then we were hesitating. We were choosing between “Tal”
and “Capablanca” – look at the preliminary notes for this game.
As a practical player striving for victory, start by searching for a move that lines up with “Tal.”
As we know, Tal preferred the “Tal” in both the first case and the second. There followed, respectively, 14.dxe6 and
29.Qb7+. Two positions, two moves, and two sacrifices!
Pay attention! I’m about to ask a question of colossal practical significance. Ready?
What do we do in positions that are complicated for us, if after totalling up the algorithm we are not sure whether to go
with “Tal” or “Capablanca”?
Answer: Approaching the situation not as a great scholar, but as a practical player striving for victory, start with the
search for a move that lines up with “Tal.”
Why?
Because otherwise you won’t find 29.Rb7!, missing an instant win! Is that what you want?
In order to become an outstanding chessplayer, it is necessary to be born “with a clot of dried blood in your right hand”
[rumored birth of Genghis Khan – Tr.]. If you don’t, then you must acquire the psychology of a victorious warrior.
That’s good. And it’s absolutely necessary in a struggle where there’s no room for fear...
We are nearing the conclusion of this chapter. Serving as a bridge between the Tal and Capablanca algorithms, the last
three examples will help to lead us into the next chapter.
As we know, the algorithms of Tal and Capablanca are neighbors (see the Algorithm Drift Chart at the end of Chapter 1
or in Chapter 6). They are friends, not enemies. They have something in common. But there are differences between
them, too.
Here the question comes in: what’s the main difference between these two algorithms?
It’s their attitude toward the safety factor. To wit: with the “Tal,” it’s “+”, while with the “Capablanca,” it’s “=”.
Of course, this answer should not be considered absolutely reliable. It’s not that easy. The time will come when we will
learn the nuances of the relationship between these two algorithms. But for us, for now, it’s important for our answer to
be correct in principle. And that’s the main thing.
3) an overwhelming advantage in the safety factor (Black’s king is not only much more poorly placed, it’s also almost
immobile; in addition, White has an extra bishop and knight inside the rectangle f1-f8-h8-h1).
As we can see, the first and second factors point in the same direction, calling us to decisive action. Our diagnosis is the
Tal Algorithm!
21.e6
There’s nothing extraordinary or miraculous about this move. Rejecting the possibility of liquidating the a1-knight, Tal
instead sacrifices a pawn. White’s intention is obvious: to retreat the bishop (to c3, say) and then occupy the vacated
square with the knight. After this, Black will hardly be able to hold his position together.
21...Qxe6
“21...fxe6 22.Qd3 (22.Rxa1? e5!) 22...Rxf6 was dangerous for Black, but by 21...Bxe6 22.Bxg5 Qc5! 23.Nf6+ Kh8
24.Bxh6 Rfd8 25.Qf4 Qc4! he could have forced the exchange of queens: 26.Rxa1 (not 26.Qg5? because of 26...Qxf1+
27.Kxf1 Bc4+) 26...Qxf4 27.Bxf4 Bxa2 28.Be5 with a rather unclear ending” (Kasparov).
However, we can improve on Kasparov’s analysis. 24.Be3 (instead of 24.Bxh6) leads to an edge for White. Something
like 24...Qf5 25.Bd4 Qg6 26.Rxa1 Rfd8 27.h4 (Rybka).
Black also stands worse after 22...Qb5 (instead of 22...Qc5). Rybka rates it ² after 23.Nf6+ Kg7 24.Qc3 Kg6 25.Bh4 h5
26.Qg3+ Kh6 27.Rxa1.
Back to the game:
22.h3
This modest pawn move was pure genius. By preventing 22...Qg4, Tal plays – strategically! The pawn advance
enhances White’s whole position. Believe me, this pawn nudge is not the last of the quiet moves. There will be a g-
pawn push, too!
Is Tal ranging outside the algorithm named for him?
Probably not. But, on the other hand, consider this what you will, dear reader. All I know is, Tal was a chess magician!
22...Qf5
22...b6, with the idea of ...Bc8-b7 and ...Qe6-d5, was stronger (Kasparov).
25.g4
In total compliance with... the second point of the strategic (Capablanca) algorithm. The line between “Tal” and
“Capablanca” gets thinner, then disappears. Harmony!
25...Rac8 26.Bc3
Optimal deployment of the pieces, primed for the coming open and direct attacks on material targets – Point 2 of the Tal
Algorithm. In fact, this is also the first point of the Capablanca Algorithm!
Later on we will see that there are no contradictions between these two algorithms – none at all!
Black is worse, but his position still hangs together. After 29...Bc4 or, let’s say, 29...Rd8, Rybka sees nothing forcing.
29...Kh7
31.Nd7+ Kg8 32.Qb8+ Kg7 33.Qf8+ Kh7 34.Ne5+- is even stronger (Kasparov).
31...Kh7
32.Ne7 1-0
3) a great “+” in the third factor, since the squares f7, f8, and g7, located right next to the black king, are attacked by the
white queen and rooks; moreover, White’s force superiority in the attacking sector – that is, in the rectangle f1-f8-h8-h1
– is evident.
Or, to put it briefly (translating from spoken language to mathematical language): m < 1, p ~1.31 > tcr = 1.25, large “+”.
This diagnosis can mean only one thing: the position requires the Tal Algorithm, because the first three parameters all
“vote” for this bloody approach.
What happened in the game?
A sacrifice! (Tal turns down a repetition by 28.Qe6+ Bf7 29.Qf5 Bg6 30.Qe6+.) There followed:
28.Rxg7+
28...Kxg7 29.Bh6+
...then a bishop...
29...Kxh6 30.Qxf8+
...but now White has won a rook. All the time, check, check, check... Forcing play: the Tal Algorithm at its zenith!
30...Kg5 31.bxc4
White nets a knight...
31...bxc4
...for a pawn.
Let’s do an interim review of this truly uncompromising and gory encounter. The question is: how are we doing in terms
of material?
Answer: status quo! It’s easy to see that White has succeeded in losing a rook, a bishop, and the b-pawn. In those same
four moves, though. he has acquired the black g7-pawn, plus the rook and the c4-knight.
A combination? Yes, yes, yes! Beyond the shadow of a doubt!
A trivial exchange of a bishop for a knight? Yes?! Or yes!? Or...
We will not torture ourselves. We will not swim pointlessly amid the flotsam of inevitably imprecise definitions. We
simply say: the combinations in this game are exchange combinations. Combinations with a clear foretaste of the
Capablanca Algorithm.
Tal’s next move involves a pawn...
32.g3
The first in a series of moves to improve the position. For now, he will do it with his g-pawn. A move later – with the h-
pawn. And then, with the king himself!
32...Be4 33.h4+
Check – be it merely the very humblest check from a pawn – is always an open attack on Target No. 1.
“Tal”? Unquestionably, yes.
“Capablanca?” Yes?! Or yes!?
33...Kg4
34.Kh2
An exceptionally strong and – just as important – very beautiful move. Tal improves his king’s position. In other words
(excuse me: I’m getting too far ahead), Tal clearly is playing strategically. He plays in perfect harmony with the first
element of the Capablanca Algorithm.
And that’s not all. Tal, though playing according to “Capablanca,” stays faithful to “Tal”! Indeed, after the game move,
White suddenly threatens mate in two with 35.Qf4+ Kh5 (Black’s king can no longer move to h3!) 36.Qg5#.
Harmony! “Tal” and “Capablanca” are wartime companions, co-workers...
34...Bf5 35.Qf6
“Continuing the pursuit of the king. After the greedy 35.f3+?! Kxf3 36.Qxf5+ Ke3 37.Qxh7 (37.Qxc2?? Rd2+) 37...c3
the passed c-pawn would have insured Black against any troubles” (Kasparov).
“36...Bd3!? 37.Qf6 (necessary, otherwise the black king breaks through to f3) 37...Bf5... would have given an
immediate draw” (Kasparov).
Tal recommended 38...Kxf2. After 39.Qxc2+ Kf3, “White cannot escape perpetual check.”
A mistake. “Black gets an easy draw with 41...Re6! 42.e4 c3, building a fortress” (Tal). But now White gets winning
chances.
The continuation was –
“Diagnosis”: the Tal Algorithm, because even a small “+” quite often has great significance...
So how did things go?
In the game, there was an open attack on the knight! There followed
19.c4
The alternative was 19.Qg3. Now 19...0-0 is impossible due to 20.Bxh6. There remain only 19...g6 and 19...g5, which
are easily seen to weaken his king’s wing.
20...Nd5
Of course 20...Qxd4+ is bad, since after 21.Be3 Black comes out a piece down.
21.c6
No doubt Tal visualized this sacrifice before making his nineteenth move.
21...bxc6 22.Ba3
22...f5
23.exf6
With this move, White tears apart the opposing king’s shelter.
23...gxf6
Pay attention: here comes a very important strategic maxim! This is it: following an exchange (a strategic exchange), the
attacking side must at once reorganize his pieces optimally. I repeat: at once!
This is a maxim that will sound a thousand times over.
What does Tal do?
Without betraying himself, Tal plays according to “Capablanca.” Like this:
24.Rb3
24...Kd8
Winning immediately. Black has no reasonable defense against the checks on b8 and g8.
26...Nb6 27.Rxb6 axb6 28.Qg8+ Qe8 29.Qxh7 Rxa3,
Following an exchange (a strategic exchange), the attacking side must at once reorganize his pieces
optimally.
We took stock of our first acquaintance with Tal’s Algorithm on the “neutral ground” between Games 3 and 4.
Remember that it was precisely there that we defined the most important aspects of our further progress:
There, we promised to:
1) flesh out all the elements of the Tal Algorithm, insofar as possible, with the richest chess substance;
2) extend the scale of material chess targets;
3) raise the status of the first three factors of the chess position.
Now, in closing this chapter, we can confidently state that we have “fleshed out,” we have “extended,” and we have
partially “raised the status.” We made real progress. In addition, we have genuinely excellent prospects.
The first aspect: We have managed to review the entire right side of the chess spectrum, which is the whole Tal Zone
(see the Algorithm Drift Chart). We did this so well that we managed also to enter into “Capablanca” territory.
It’s totally obvious that the four elements of the Tal Algorithm have shown their reliability and effectiveness. They
work.
I have combined the first two elements of the algorithm, “shooting” through the occupied and unoccupied squares of the
chessboard. “Tal” primarily involves attacks on occupied squares, whereas “Capablanca” features attacks on
unoccupied squares. The important thing is that these algorithms intersect.
“Tal” primarily involves attacks on occupied squares, whereas “Capablanca” features attacks on
unoccupied squares.
The important thing is that these algorithms intersect.
In attacking the occupied squares of the chessboard, we also thereby attack its empty squares. In this, the number of
empty squares is always greater than the number of occupied ones. And this means that the Tal Algorithm is simply
fated to morph, sooner or later, into the Capablanca...
I have also joined in tandem the third and fourth elements of the Tal Algorithm. The name of the union is an “unequal
exchange.” In such an exchange, pieces of unequal value come off the board. When attacking, we do not simply
sacrifice chess material – we sacrifice material in order to recover it later with interest.
With the “Capablanca,” the unequal exchange collapses and turns into a strategic trade...
Second aspect: I tied it to the problem of extending the attacking scale (from attacking material targets to attacking
empty squares). I will say at the outset: extending the scale is easy. But then we need to be very careful!
Why?
Because the chess values of empty squares vary in their importance from the vanishingly small to the boundlessly great.
And so here’s why, like it or not, we are forced to share these empty squares between the two algorithms. Part of them
goes to “Tal,” the rest to “Capablanca.”
“Tal” is responsible for attacks on empty squares that feature direct attacks on material targets. “Capablanca” is for the
attack on empty squares, conceptually distinct from the attacked non-empty squares. Of course, in practical chess we
are dealing with an unimaginably great number of intermediate conditions...
The Capablanca Algorithm will stand for the attack on unoccupied squares whose chess value is less than the chess
value of a pawn (i.e., less than 1). These small values can and do add up to more than the value of a pawn, and even of
a piece. I repeat: from here on out, “Capablanca” stands for attacking empty squares of modest chess value. Make it so!
In this way, we formulate a complete priority scale for attacking the occupied and unoccupied squares of the
chessboard: king, queen, rook, bishop or knight, pawn, empty square. The chess values to be applied to them are:
infinity, 9, 5, 3-3.5, 1, and a number between 0 and 1 (where 1 represents the value of a pawn).
The values of empty squares in the Tal Algorithm are randomly scattered from the infinitely large down to 1. Empty
squares in the Capablanca Algorithm are limited to the range from 0 to 1.
The safety factor cannot be boiled down to a single number. There are positions in which it’s impossible to
find “a” best move. In such positions there can be two or more “strongest moves;” then you have a choice.
Third aspect: I linked it with the problem of raising the status of the first three factors of a chess position. I would prefer
these factors to be fundamental, which is to say, unobjectionable from a scientific standpoint...
What do we have? Are these first three factors of the chess position fundamental or not?
Only three factors have withstood the test of time: the material factor, the factor of chess time, and the space factor.
These three factors – in clear or disguised form – hop from one theory of chess play to another. They have hopped thus
for over a hundred years, from Steinitz’s time to the present!
On these three key elements – on these three fundamental factors of the chess position – there rested, rests now, and
shall forever rest the whole of chess science. This includes the theory of chess play that I am teaching.
The three aforementioned factors are but a hunger ration. I repeat: a hunger ration for any theory of chess play. It’s the
bare minimum – there’s nothing smaller. A larger serving is owned by the great majority of chess authors...
Up to this point, we have made use of only two of the three fundamental factors – material and time. The next factor in
our model that we will discuss is the factor of safety – which is non-fundamental.
Why?
Because it stands at the juncture of the three fundamental factors – material, time, and space.
The material factor, the time factor, and the space factor have given birth to the safety factor, each delegating to it a part
of their legitimate power. Moreover, they have delegated the biggest part. The fundamental factors have given up all of
their irrationality to the factor of safety, thereby washing themselves clean and becoming transparent. Now they are
defined by numbers, easily and together. A rational number. In any chess position!
But the safety factor cannot be defined by a single number. There are positions in which it’s impossible to find “a” best
move. In such positions there can be two or more “strongest moves;” then you have a choice. The imagination is
engaged. Chess truth slips away. It hides from us in the impenetrable jungle of irrationality.
As we conclude this chapter, I’m not afraid to remind you once again, dear reader, of the three most important aspects
of our progress. In our approach to the Tal Algorithm, we have “fleshed out,” we have “extended,” and we have “raised
the status.”
First, we have – most importantly – fleshed all four elements of the Tal Algorithm, without exception, with the very
richest kind of chess substance.
What comes next?
Later on we will complete the smooth transition from the Tal Algorithm to the Capablanca Algorithm. We will
transform the four elements of the Tal Algorithm into the three elements of the Capablanca.
Secondly, we extended the scale of material chess targets to include unoccupied squares, à la Capablanca.
Further, we have concentrated our attack on empty squares, à la Capablanca.
We – and this is in the third place – have significantly raised the status of the first three factors in any given chess
position. The first two are fundamental; the third is a blend.
All of this means that we keep moving. And our next movement is to come to the Capablanca Algorithm.
We are filled with decisiveness, absolutely certain of success. We will triumph. We shall overcome every obstacle!
The door is open. Let’s go in!
Chapter 3
Please note: Soon, we shall have dealings only with the “left-hand Tal” and “right-hand Capablanca.” We occupy a
sector of the right-hand portion of the unified spectrum of all chess attacks and defenses. I repeat: we are only
interested in the right-hand portion of the spectrum!
When attacking in “Capablanca” style, we attack the unoccupied squares of the chessboard, of small chess value. Their
value is less than that of a pawn – that is, less than 1 point.
In a strategic attack, the dominant factor is always the optimal deployment of the pieces, carried out over several moves.
These setups are almost always enhanced by pawn moves – the pawns “come up”... The struggle takes place on the
opponent’s territory.
As a rule, the three elements of the Capablanca Algorithm are applied in the following order:
1) optimal placement of our pieces on squares suitable for a strategic attack on vacant squares;
2) pawn advances;
3) exchanges.
Short version: in order to play “the Capablanca,” we have to:
1) put our pieces on their best squares;
2) advance our pawns;
3) seek exchanges.
The elements we have just listed require us to do some detailed decoding. We need to fill out the elements of the
algorithm with rich content. In this task, Capablanca himself will help
us!
Our goal is obvious – to dissect Capablanca’s method for searching for the best move. As we investigate the games of
the third World Champion, we will try to generate proofs of the “algebra” of this method. We will perform our task
humbly, with reverence and the greatest zeal...
Are you ready?
3) approximate equality in safety, as both kings are securely and tightly tucked away, and White’s barely noticeable
pressure on the opponent’s kingside (the e6-pawn and the “X-rays” emanating from White’s queen on c3) are fully
compensated for by Black’s “extra” rook in the rectangle f1-f7-h7-h1, where both kings are found.
Attention! Two surprises await us in this game. Those are: a) m > 1; and b) approximate equality with respect to the
third factor of the position. We haven’t had anything like this before (you can check this out). We’ve had examples
where m < 1 or m = 1, and there was (in all the examples) a steady “plus” with respect to king safety.
Since, as we know, when m < 1 the first factor impels us to the “Tal,” then if m > 1 it simply must push us in the
opposite direction. It impels us toward the “Petrosian.”
And now, for the third factor. Short and simple: it’s neutral. In this example, it’s equidistant from the algorithms of Tal
and Petrosian. It’s neither “plus” nor “minus.” More precisely, its pressure on us to “add up” the algorithm is exactly
nonexistent.
Diagnosis: both “Tal” and “Capablanca” at the same time.
Detailed commentary on both the diagnosis and the two aforementioned surprises will come later, in the postscript. I
add that a micro-correction to our diagnosis is possible, since at this stage it’s only preliminary.
This multiple diagnosis is the primary reason for our difficulties. These problems are most closely bound up with the
choice of attacking targets. As is known, they can be not just the pieces and pawns of our opponent, but also the empty
(that is, free of chessmen) squares of the board.
So what can we do? Which objectives do we attack? Empty squares? Material targets?
The answers will come to us from considerations of space!
Say that again?
Yes, from considerations connected to how closely packed the white and black pieces are around their respective kings
and pawns. In the current position, White’s forces are scattered, sprawling from the a- to the h-files and from the first to
the sixth rank. In that 8 × 6 rectangle, there are 48 squares altogether.
Setting aside the white queen, the rooks on e1 and e5, and the light-squared bishop, we are left with the king and seven
pawns – eight chessmen in all.
Now let’s do the same thing with the black pieces: set aside the black queen, the rooks on e8 and f6, and the bishop at
c8. That leaves king and six pawns, or seven chessmen.
And in what size rectangle? In an 8 × 3 rectangle, bounded by the lines a7-a5-h5-h7, where there are 24 squares.
This means that the density of White’s king-and-pawn position is considerably lower than the density of Black’s: 8/48
<< 7/24. Black is more compact.
This means that an immediate attack on material targets is impossible: one cannot break through Black’s defenses with
such widely dispersed forces.
White needs patience and time in order to prepare for the inevitable – but, as will be seen later, victorious – tactical
battle. He must collect his troops and deploy them optimally. Of course, his pieces stand “poorly.”
How, exactly?
I immediately declare a total mobilization, and then inspect my valiant troops. I call upon the senior and the junior
officers of the ivory army. I summon them strictly according to rank.
Of course, I start with the general on g1. He is – as is proper in the order of battle – protected in a secure bunker. In
time, I will open an escape hatch to save him from random shots along the first rank.
The queen: No complaints here; she is in the thick of the action. Through the palisade formed by the rooks, the queen
gazes steadily at the black king.
Next are the rooks. Unmoving, like glowering bulldogs, they are set to watch the living body at e6. When I give the
order for it to perish, at that moment the rooks must show what they’re capable of. They will do their thing along the e-
file. No doubt about it: the rooks are well placed.
Now the bishop, the one on d3. The bishop is... an eyesore! He’s not quite a deserter, but in the best of cases an
unwilling combatant! He bites on granite at f5. Useless lout! I call out threateningly: “To the front line!”
25.Be2
Capablanca redeploys his passive piece so as to aim at d5. However, this was not the only possible decision. 25.Bc2,
with the idea of 26.Ba4, doesn’t look bad either.
25...Bxe6
Fritz prefers 25...Qd6. But here, too, after 26.Rd1 Qc7 27.Bf3 it gives a ±, that is, a great advantage. One possible line
is 27...Rexe6 28.Red5 Re8 29.Rd6 Ree6 30.Rd8 … 31.g4, with highly dangerous threats.
Now let’s compare these parameters with those of our starting position.
What do we see? An obvious change in trajectory towards the Tal Algorithm! The changes in “m” and “t” tilt us that
way!
What does Capablanca do?
28.Qe3
He attacks the newly established target: the enemy bishop. All the more so, in that the black king has vacated the
dangerous a1-h8 diagonal.
28...Re7
What now?
29.Qh6
29...Kg8
Here 30.Qg5 or 30.Qe3, hitting the strongest and most accessible targets, would be silly because of 30...Kf7, and
because (as the stronger side) we don’t want to repeat the position. If White doesn’t need a draw, then given that the
first part of the strategic algorithm has been accomplished, he must activate the second part!
We coolly establish the fact that, in the position after 29...Kg8, the white king and six pawns stand within the rectangle
a1-a4-h4-h1, sized 8 × 4 = 32. This means that the density of the white position involving king and pawns is 7/32.
What about Black? King and six pawns in the rectangle a5-a8-h8-h5; the area of the rectangle is 8 × 4 = 32 squares. In
other words, the density of the black position of king and pawns is 7/32.
7/32 = 7/32.
After 29...Kg8, Black’s defense becomes significantly more porous – that is, of limited effectiveness – and the density
of the opposing sides is comparable. This means that the aggressor (White) is entitled to a decisive assault.
30.h4
In complete accordance with the second part of the strategic algorithm. The whole force of Capablanca’s position is
aimed at the potential energy of this flank pawn. At the same time (see above), we make an escape square for the king.
30...a6
31.h5
Is this an open attack on the g6-pawn (the first part of the Tal Algorithm!?), or an “innocent” play for trades (the second
and third part of the Capablanca Algorithm)?
This dynamic phase of the battle is relatively short. If 33...Rexe6, then 34.Rxe6 Rxe6 35.Qxg6+, and Black loses a
piece.
Black resigned. 1-0
Note that, in the final position, White’s king and five pawns stand within the rectangle a1-a4-g4-g1 (7 × 4 = 28),
meaning that the density of White’s position, for the king and pawns, is 6/28.
Meanwhile, Black is less compact: his king-and-pawn density is 6/35.
6/28 > 6/35. This is one of the reasons for Black’s swift collapse, the reason that flows from “Capablanca.”
But there is also a reason based on “Tal.” Judge for yourself: in the rectangle f1-f8-h8-h1, where the two kings lie,
White enjoys material superiority. Additionally, in the final position the white king is safe, whereas the black king is
(pardon the expression) stark naked. And one more problem: of the five vacant squares surrounding the black king, four
are controlled by the white queen. Almost total paralysis!
And the last reason before the postscript. Note that in nine moves of the attack, White improved more than just the
bishop on d3 to d5. He has also improved the queen, from c3 to h6.
No further comments for now.
The “Tal” is the opposite of the “Petrosian,” and the “Capablanca” is always our golden mean in the
spectrum of attack and defense.
Postscript
To begin with, our first surprise – the first factor of the position and its parameter, “m.” In the starting position, our
factor is m > 1.
Question: how does the “m” parameter affect the process for selecting the right search algorithm? I will try to make the
answer short and clear.
Answer:
1) if m < 1, then we shift to the right – that is, from the Petrosian Algorithm to the Tal Algorithm; in this case, the
important thing is for our psychological state to be that of a hungry chess wolf (an elevated level of chess
aggressiveness);
2) if m = 1, then we “stand in place,” and our psychological state is neutral (a medium level of aggressiveness);
3) if m > 1, we shift to the left – that is, from the Tal Algorithm to the Petrosian – and our attitude is invariably one of a
sated and almost always peaceably inclined guard dog, protecting the home and master (a lowered level of
aggressiveness).
Why are we like this?
Short answer: because the “Tal” is the opposite of the “Petrosian,” and because the “Capablanca” is always our golden
mean in the spectrum of attack and defense (see the Algorithm Drift Chart).
All these recommendations – even requirements – of the “m” parameter will be considered universal. We will consider
that the recommendations and requirements work every time. We will use them in every calculation of the strongest-
move algorithm, in every position, without exception.
Our second surprise (returning to the starting position above) is the third factor in the position, that of safety. Parity!
Previously, we’d had no cases of parity in this factor. We had ten “Tals” and ten “pluses” in terms of the third factor.
Was this a coincidence?
From here on out, we will operate on the assumption that this was not an accidental, but a logical outcome. In this way,
we acknowledge the decisive role of the third factor in the calculation of our algorithm. As of today, a “+” in the safety
factor is associated with a strong drift to the right, toward the Tal Algorithm. And the greater the “+”, the stronger the
drift!
Of course, if the safety factor is negative, then the situation is reversed. A minus sign (“-”) implies a drift to the left, to
the Petrosian Algorithm.
What’s left for the Capablanca Algorithm, then? The answer is pretty obvious – parity in the third factor. “Capablanca”
covers situations where neither king is in danger. It is the zero of the safety factor that is encountered between the “+”
and the “-”.
Naturally, all of our assumptions remain just that, for now. But, I have to say, they are very close to the truth.
The safety factor is very fickle. It explodes at the lightest touch, and I ask the reader to treat it delicately.
A “+” in the safety factor is associated with a strong drift to the right, toward the Tal Algorithm.
13.Bxf6
13...Nxf6 14.Ne4
Capablanca offers a trade on e4 (the more trades, the better for him). In other words, Capablanca plays “according to
Capablanca.” Strategically flawless play!
On the other hand, Capablanca is also playing “like Tal.” He is playing in a most crude manner, attacking both the c7-
pawn (a one-move threat) and the h7-pawn (a two-move threat).
Algorithmic harmony!
Or 15...b3 16.Bxh7+ Kh8 17.Qd3 g6 18.Bxg6 fxg6 19.Qxg6, with an attack (see José Raúl Capablanca, vol. 1, 1997,
edited by S. Solovyov). True, Rybka still holds the position together somehow after 19...Bd6 20.Ng5 Re7, as well as
after 19...Bb4+ 20.Ke2 Re7.
16.Qxe4
Now White intends to get a rook to the empty spot c6. Such play is called “playing in the gaps.”
16...c5
Practically forced.
Black has avoided any unpleasantness on the c-file, but he’s jumped from the frying pan into the fire!
19.Ng5
A typical, or, if you will allow me, a “100% Tal” in Capablanca’s hands. The king’s wing is weakened, and there,
following the line of least resistance, is where the stronger side is sending his streams of fresh energy. I add that here
we are simply obligated to play “in Tal style,” if we want to play as strongly as possible.
Why?
Because, within the f1-f8-h8-h1 rectangle, the black king faces White’s “extra” knight and rook. Besides, for now, no
one has succeeded in removing the white queen’s fundamental pressure against the h7-pawn. And this pawn is located
in the immediate vicinity of the black king!
19...h6
An error – which Capablanca fails to exploit. Black had to reply 19...g6, when there is probably no win for White.
20.Qf3, 20.Qf4, and 20.Qh4 are all tempting. Everywhere we get variations, variations, variations...
Analyze the position after 19...g6 on your own. And, this time, do use a computer!
20.Qh7+
20...Kf8 21.Qh8+
The forcing play is over, and White must busy himself with the proper positioning of his still inactive pieces. Point 1 of
the strategic (?!) algorithm of the search for the strongest move? Or is it still “Tal”?
Answer: both “Tal” and “Capablanca” at the same time!
24.Ke2
Capablanca improves his king. The purpose is clear: White intends also to bring up the rook from h1 to c4.
By improving the king, Capablanca substantially improves the mobility of his pieces (the king and the rooks). But the
chief reason it’s an improvement has nothing to do with dynamic considerations. It has to do with a static element: with
the text move, White sharply increases his compactness (how densely packed together the king and pawns are), from
7/32 to 7/24. Check it out!
24...Rac8
Computer programs look very favorably on this move. But they are no less favorable to 24...Kc6. A rough variation
goes 25.Qf4 Ra7 26.h4 f5 27.h5, with a clear edge.
25.Rc4 Kc6
But this king move is most likely unfortunate. Rybka thinks 25...Rc7 is better, and if 26.h4, then 26...Kd7, skipping off
to e8. A rook move, 26.Rhc1, may be even stronger than pushing the h-pawn: White is better in all variations.
26.Rhc1 Kb6
27.h4
All of the white pieces are placed ideally. That’s “One.” “Two” is the pawn advance!
27...f5
“I expected the reply 27...Rc7 and had readied the following variation: 28.h5 Rec8 29.h6 Bd6 30.Qxa5+ Kxa5 31.Rxc7
Rxc7 (31...Bxc7 32.Rc6) 32.Rxc7 Bxc7 (Black comes out of the fight with a material advantage, but his king is on the
queenside, while the outcome must take place on the kingside. And in endgames, when almost no pieces remain on the
board, except for the pawns, the king’s absence from the focus of the struggle is usually fatal. White must strive to
queen his pawn before Black’s king can get over.) 33.f4 Bd8 34.g4 Bf6 35.g5 Bh8 36.e4 Kb6 37.f5 exf5 38.exf5 Kc5
39.g6 fxg6 40.fxg6, and Black has no defense against 41.g7” (Capablanca).
Your task is to find the win after 34...Be7 (which does exist and is not too complicated). This move is stronger than
moving the bishop to f6.
One more task: wouldn’t it be simpler to play the “Tal-like” 32.h7 (instead of 32.Rxc7)?
28.Qg7
30.Rxc5
17.c4
A paradoxical move which confounds the Cuban genius’s many fans. It overturns false values, it is stunning. This move
is possible only because White has a clear space advantage (that pawn on f5!). Black is suffocating: he suffers from a
serious lack of suitable squares for his shiftless pieces. This is why it’s good for Capablanca to refrain from trading
bishops.
And now, let’s run some numbers. We set this position a single question: what sort of ∆k do we have?
The Fourth Parameter: (∆k), or the measure of one side’s superiority over the other in the factor of
compactness.
17...c5
18.g4
White exploits his opponent’s hesitation (failing to play 17...b5) at once and in exemplary fashion. A target – the f6-
pawn (see move 20) – is attacked.
One pawn move is answered by another. Symmetry! Among other things, he threatens the crushing 21.g6+. Capablanca
playing in “Tal” style? Yes!
Why? Calculate the mobility of the white and black pieces following 19...b5. White has moved into the Tal Zone on the
proper basis. Woe unto him who is weak!
As we know, the sources of excess energy of a chess attack spread themselves according to value: K, Q, R, B or N, §,
and empty square (or squares). Next in the order of attack is the h7-pawn, which covers the black king.
22...bxc4
If 22...Bh5 (Fritz’s first line), then 23.Rg3 h6 24.cxb5, with advantage to White, is very possible – the g5-knight is
“poisoned.” Prove it!
23.Nxh7
We attack the bishop, thereby attacking the king: White threatens 26.Bxg8 and 27.Qxh7#.
25...Rf7
“Incontrovertibly, the move 25...Nf7 looks better. Nevertheless, White could continue with either 26.Kh1 and then
27.Rg1, or 26.f6! g5 27.Qh5 Nd8 28.Qh6! Rf7 29.Bxg5 Qf8 30.Kh1! Qxh6 31.Bxh6 Rxf6 (31...Nxf6 32.Rg1) 32.Rg1,
winning by a rapid assault” (Capablanca).
The computer brings a correction to this analysis. It goes more or less like this: 25...Nf7 26.f6 g5 27.Qh5 Nd8 28.Qxg5
(instead of Capablanca’s line) 28...Rf7 29.Bxf7 Qxf7 30.Rf1 Ne6 31.Qh4 (… Bh6, Rg3+-) 31...Rd8 32.Bh6 Rd7
33.Kh1+- (Fritz).
The task now is to check Fritz. And if necessary, to extend its analysis!
26.Kh1
26...b5
If 26...d5 27.Bxd5 Rd7 (pointed out by Fritz), then 28.Bc4 and White must win.
27.Bd5
27.Bxf7 is even stronger. After 27...Qxf7 (27...Nxf7 28.f6 g5 29.Bxg5+-) 28.Bg5 Nc6 29.f6, Black is helpless.
29.Bg5
White’s last piece has occupied an ideal attacking stance. This means that Capablanca, while attacking in “Tal” style,
still heeds the strategic algorithm. Capablanca plays in the style of “Capablanca”!
29...Raf7
30.b3
Point 2 of the strategic algorithm in action. Capa advances this pawn in order to secure the a-pawn’s march to the
queening square. In other words, playing strategically, Capablanca first of all considers how to improve his position. He
doesn’t look for concrete, forcing solutions till his position has been improved to the utmost.
What is allowed to the gods is denied to mere mortals. But we are permitted a method which has already worked quite
well.
In the diagram position, m < 1; t = 39/21 = ~1.86, which is >> tcr = 1.25; a large “+” in the safety factor; and
unquestionably “Tal,” “Tal,” and again, “Tal”!
The first three lines on the Fritz screen are 30.Bxf6, 30.Rhg3, and 30.Bxf7 – all three are rated “+-”.
One absolutely crushing variation: 30.Bxf6 Rxf6 31.Rhg3 Qf8 32.Rxg7 Qxg7 33.Rxg6, etc. Examine it on your own.
It’s truly surprising (isn’t it miraculous?) that, while improving his position, Capablanca doesn’t let the win slip even for
a minute. Clearly he is in no rush, and Caissa (the goddess of chess) will bless him for it!
30...Qf8 31.a4
34...Nb4
Capablanca first of all considers how to improve his position. He doesn’t look for concrete, forcing solutions
till his position has been improved to the utmost.
35.Bxf6
35...Nxd5 36.Bxg7+
And this check – open attack on the king – is the first element.
and Black resigned soon thereafter. A curious finish would be 39...Nf4 40.Qf6#.
This afterword to the game I begin with a very important definition. We must define the fifth and last of our factors of
the chess position. This is the factor of spatial expansion. This factor and its corresponding parameter define the degree
of “enhancement” or improvement in the position.
The Fifth Parameter: ∆(move), or the factor of spatial expansion.
We will assign each of the opponent’s pieces (see the diagram above after 34...Nb4) a whole number, equal to the
number of the rank on which the given piece stands.
For example, we assign the white king the number 1, since it stands on the first rank. The same number, 1, is assigned to
the white rook on g1. The h2-pawn gets a 2, and so on, right up to the “elevated” pawn at a6, which gets a 6.
We add these together (1 + 1 + 2 + ... + 6) and then count the number of white chessmen. We get 39 and 11. The
number 39/11 will establish for us the white position’s center of gravity. An important preliminary result!
And what about Black? Let’s tally it up: 1 (starting from the eighth rank) + 1 + 1 + ... + 5, and then we calculate the
total number of chessmen. That comes to 28 and 11, meaning that the black position’s center of gravity corresponds to
28/11. An equally important preliminary result!
Now we calculate the arithmetical difference between the numbers that define the center of gravity of the white and
black armies. We denote this value as follows: ∆(34...Nb4). Important: subtract the opponent’s number from the
number for the player whose turn it is to move. In this case, this works out to 39/11 – 28/11 = 11/11 = 1.00.
Or, in short, ∆(34...Nb4) = 1.00.
We calculate this parameter precisely to two decimal places; the parameter itself will be indicated by the symbol
“∆(move)”.
The parameter ∆(move) is the fifth and last of the parameters of a chess position. It represents the amount of
“improvement” in the position in terms of space. It is an objective measurement of this expansion – of increased chess
space.
The factors of expansion and of compactness are the two aspects of the dual-natured space factor; ∆k and ∆(move)
complement each other.
Now for a few words about how these two parameters influence the process of selecting the search algorithm (see the
Algorithm Drift Chart in Chapter 6):
a) when ∆k > 0, we drift to the right;
b) when ∆k < 0, we drift to the left;
c) when ∆(move) > 0, we drift to the left;
d) when ∆(move) < 0, we drift to the right.
Naturally, if ∆k = 0 or ∆(move) = 0, then we stay put.
Question: if ∆k > 0, why do we drift to the right – that is, from the Petrosian Algorithm towards the Tal Algorithm? To
put it another way: if ∆k > 0 (if we are more compact), why should it be good for us to play more aggressively?
I have already answered these questions – at least in part – in the notes to Capablanca’s moves 13 and 14 in the
Schroeder game. Now I elaborate on those considerations and make a general observation.
The more ∆k > 0 (that is, the more compact we are), the more our position resembles a coiled spring. Set it loose, and it
will jump open with great force! This “spring” is ready to hit the enemy hard at any moment. It is clearly working in
“Tal” style!
Nature abhors a vacuum. The spring will keep uncoiling until it meets sufficient resistance. Or: space expansion in chess
stops only when the compactness of White’s position equals that of Black’s. Expansion is halted when the force of
action is equal to the strength of the reaction: Newton’s Third Law!
This is all true, but under one obligatory condition: when parity exists in all of the other four parameters of the chess
position!
Naturally, when ∆k < 0 our reasoning works in the opposite direction. In other words, in this case the parameter ∆k
impels us, not toward “Tal,” but toward “Petrosian.” Symmetry!
Next step: the fifth parameter. Why, if ∆(move) > 0, do we drift leftward – that is, from the Tal Algorithm to the
Petrosian Algorithm?
The answer must be sought in the peculiarities of the very nature of the pawns. These genuinely most amazing
chessmen can, in the course of the game, turn into other pieces – of greater value.
As they improve their position, pawns get uniformly “heavier” – their potential energy increases. They rise in
accordance with their interaction with the approaching promotion squares on the eighth (or first) rank. And of course,
the further the pawn is “improved,” the greater the ∆(move) of the position – that is, the more ∆(move), the “heavier”
the position grows. The parameter ∆(move) is, to some extent, equivalent to the “m” parameter!
This is why, if ∆(move) > 0, we drift towards the Petrosian Algorithm. In fact, we have something to lose here. Which
means that we also have something to defend.
Of course, looking at it from the highest considerations of symmetry, we may confidently observe that, when ∆(move) <
0, the parameter ∆(move) works in the direction, not of “Petrosian,” but of “Tal.” We drift rightward, we attack.
Naturally, we only attack where the other parameters of the position allow it...
Concluding the afterword to this remarkable game, we return to its “sources” – see the starting position. We refine our
“verdict”...
What do we have? ∆k < 0 and ∆(16...Bf7) = 0.40 > 0. Verify this!
Have you verified it?
The fourth and fifth parameters of our position make us re-examine our preliminary verdict. We must soften it
somewhat: not “Tal” and “Petrosian” simultaneously, but the right-sided “Capablanca,” standing off to the side, as the
fourth and fifth parameters nudge it leftward.
This is why Capablanca was obliged to play the “Capablanca.” And why he pushed the c-pawn...
As we saw, Capablanca, on the attack in the game with Molina, kept “elevating” his position. The idea of space
expansion is perhaps one of the dominant elements of his genius, the other one being the idea of compactness.
We strive for error-free play in all possible chess positions, without exception. Our goal is unattainable, yet there is an
eternal desire to reach it.
Our ultimate objective is to have a universal method for seeking the strongest chess move in all possible positions. The
goal is unique – more accurately, triply unique, its facets being the fundamental algorithms of Tal, Capablanca, and
Petrosian.
In the here and now, our concrete goal is the next game by Capablanca.
Ready?
3) a small but stable “plus” in the third factor of the position, as every single one of the enemy pieces is paralyzed either
partially or completely (including the poor king on e7).
Our diagnosis is unanimous: “Tal.”
However, in our “Tal” we cannot fail but to notice one very uncomfortable peculiarity of the position: we are very much
behind our opponent in terms of compactness. This means that there is no chance to break immediately into Black’s
camp. The attack must be prepared carefully.
39.b6
One alternative to the text is 39.Ra1 … Rhh1, Rhb1, Qa7, bxc6+-. And if 39...Ra8, then 40.Qxa8 Rxa8 41.Rxa8 (…
42.b6+-) 41...cxb5 42.Rg8 … Rg7, Rxh7+- (see José Raúl Capablanca, vol. 2, 1997, ed. by S. Solovyov).
39.b6 is a “Tal” move: open attack on the queen!
39.Ra1 would also be a “Tal” move. But at the same time it’s also a “Capablanca” move! How?
Because the second point of the Tal Algorithm corresponds to the first point of the Capablanca Algorithm.
Either one wins. Capablanca chose the more beautiful solution!
39...Qb8 40.Ra1
White intends to post a rook on a4 and the queen behind it, completely paralyzing all of the black pieces and threatening
Ra8.
47.Kg3
Before bringing his knight to a5 (via Nf3-d2-b3-a5), Capablanca gives his king a good post. The place for the king is on
c3 or g2, as necessary prophylaxis against ...Nf7xg5 and ...Qb8-h2. But why rush?
Capablanca is enjoying his position. In his fascination, he temporarily “forgets” about winning the game by
maneuvering his knight to a5. It is notable that, having set the “autopilot,” Capablanca automatically sends his king in
the direction of his pawn promotion and increases his ∆(move).
Or 51...Ke7 52.Nd2, etc. A possible variation: 52...Rg8 53.Nb3 Rcf8 54.Na5 Nd8 55.Qf1 Rh8 56.Ba6 bxa6 57.Qxa6
∆58.Ra8+- (noted by Rybka).
52.Nd2
3) a micro-advantage in the safety factor: the black king is somewhat immobile, but we have a very hard time exploiting
his predicament – we’re deep into the endgame.
The verdict is unanimous: this position requires the Tal Algorithm.
It is obvious, however, that this “Tal” is not a “100% Tal”: there won’t be any genuine fireworks, because there’s little
material left on the board. And we are still behind our opponent in compactness: 4/24 << 2/6. Check it out!
Because our ∆k << 0, the fourth factor of this position impels us to the “Capablanca.” The same happens with the first
factor (m > 1). And finally, we are drawn away from the Tal Algorithm to the Capablanca Algorithm by the position
itself: our ∆(69...Nd8) > 0.
What does Capablanca do? He increases his already large ∆(move).
72.g6
Signaling an attack on the king. I add that an attack on Material Target No. 1 is combined, for Capablanca, with his
space expansion, which doesn’t let up for a second. And here the expansion is dominant, it is a goal. The mating attack
is merely a sideshow to the expansion!
72...Nb7 73.Ne8
Preparing to check with the g-pawn and restricting the the enemy knight’s mobility; now the d6 square is verboten.
73...Nd8 74.b5
Depriving Black of c6 for the knight and a6 for the bishop. Capablanca plays for restriction?! Yes or no?
Interpret this, dear reader, any way you wish! What we see here is no struggle. It’s agony for Black. He is choking, two
or three steps from total paralysis.
74...Kg8 75.g5
Capablanca is in no rush.
75...Kf8 76.g7+
30...Nf5
Black has the right to this exchange “out of order,” that is, before completing the first two phases (putting the pieces on
their best squares and then advancing the pawns). Why?
Because his position is the more compact! If your position is more compact than your opponent’s, then under the
Capablanca Algorithm it’s OK to trade off some pieces before improving the rest or advancing the pawns.
The fewer the pieces left on the board, the more chances Black has to jump into the “gaps” and establish himself there.
If your position is more compact than your opponent’s, then under the Capablanca Algorithm it’s OK to
trade off some pieces before improving the rest or advancing the pawns.
31.Nxf5+
If 31.Red2, then 31...Nxd4 32.Rxd4 Rxd4 33.cxd4 Qb5 34.Qf3 Rc1(“µ;” see José Raúl Capablanca, vol. 2).
Rybka prefers 32.Qe3 (instead of 32.Rxd4), but here too, it awards Black a clear advantage after 32...Qb3 33.Rxd4 Rxd4
34.Qxd4 b5.
31...gxf5 32.Qf3
32...Kg6
Just in case. Black protects the h-pawn and... “elevates” his position. There is no doubt that Capablanca made this move
automatically, instinctively.
We already know that faultless strategic play in superior positions is practically always linked to increasing our
∆(move). While not decreasing (and preferably increasing) our compactness.
33.Red2 Re4
36.Kg3
If 36.Rxc4 Qxc4 37.Rd4, then simplest is 37...Rxd4. Black is more compact, and trades favor him.
Sample variations:
a) 38.Qxd4 Qe2+ 39.Qf2 Qg4+ 40.Kf1 (40.Qg3 h4-+) 40...h4 41.a4 Qd1+ 42.Qe1 (42.Kg2 h3+-+) 42...Qf3+ 43.Qf2
Qd3+ 44.Qe2 Qe4, and Black must win. Find the win after 45.Qxe4 fxe4 – without the computer!
b) 38.cxd4 Qd3 39.h3 a4 40.Kh2 (White is in a pre-Zugzwang state) 40...Kg7, and so forth on the queenside – Black
should win!
36...Rcxd4
I repeat: Black is entitled to make this exchange out of turn – he is more compact.
37.cxd4
37...Qc4
Black is better, but first – as expected in a strategic attack – he improves his pieces ...
38.Kg2
38...b5 39.Kg1 b4
40.axb4 axb4
Over the last four moves, Black has transformed one compact position into another compact position – with a greater
∆(move). In the words of Alekhine, “pawn moves have the goal of removing possible objects of enemy counterplay...”
41.Kg2 Qc1
After Point 3 of the strategic algorithm there comes Point 1 – continuing on to the gain of material or (what is
practically the same thing) to Zugzwang.
42.Kg3 Qh1 43.Rd3 Re1
46...Rf1
Now on 47.Qe2, 47...Qg1+ 48.Kh3 Re1 decides; or on 47.Qg2, the prosaic 47...Rg1.
White resigned. 0-1
As a postscript to this game, here are the technical parameters of the position after 30.Re2, 32.Qf3, 40...axb4, and
46...Rf1 (see the corresponding diagrams). We are interested only in the fourth and fifth parameters:
• after 30.Re2: ∆k > 0 and ∆(30.Re2) = -0.25;
• after 32.Qf3: ∆k > 0 and ∆(32.Qf3) = 0;
• after 40...axb4: ∆k > 0 and ∆(40...axb4) ~0.89;
• after 46...Rf1: ∆k > 0 and ∆(46...Rf1) ~1.11.
Please verify my simple arithmetic. One more thing – I ask you to finish this sentence: “faultless strategic play in
superior positions is practically always connected with an increase in _____.”
18.Nf1
The first point of the strategic algorithm in action – the knight is repositioned on e3. However, 18.Nb1 … Nc3-a4
doesn’t look bad, either. Significantly, our knight on the rim is not dim. Why?
Because from a4 it attacks the poorly guarded c5-pawn. This pawn is not held by any other pawn, so it is “bad”...
18...a6
The chess engines Rybka and Fritz 11 prefer many other moves: 18...Rc6, 18...f5, 18...Nf6, 18...Qf8. Everywhere Black
is a little worse, but nothing more than that.
On the other hand, 18...a6 is more than likely just a little error.
19.Ne3
Improving (“elevating”) the knight and increasing his total mobility. Capablanca has improved his whole position –
even if only by a little.
19...Nb8
But this knight maneuver cannot be called good. Better to play 19...Ne4 or 19...Nf6 (Fritz).
20.Nd5
20...Qf8
Inching ever closer to the c5 target. Exchanges hold no terrors for White, since the opponent has the same packing
density.
23...Bxg2 24.Kxg2
26.Ndxc5
26...Qc6+ 27.Kg1 h5
Black’s “attack” (26...h5 … ...h5-h4-h3) will be beaten back easily enough. Why?
Because White is superior in the first, second, fourth, and fifth factors of the position.
28.Nd3
Or 28.Rd3 h4 29.Qd2 h3 30.f3+- (Fritz), or 28.Ne4!? h4 29.g4+- (see José Raúl Capablanca, vol. 2), or...
This is not just an exchange, it’s also a check – that is, an attack on the most important target. This is “Tal,” pure and
simple! Why?
Because, in the position after 31...Qa8, t = 46/34 = ~1.35 > tcr = 1.25.
32...Rxd8 33.Qxc7
33...Rf8 34.Nb6
In this position, m = 1, t = 1.00, “=”, ∆k = 0 and ∆(before the first move) = 0.00.
Next, we make a “projection” of the position after 17...Rac8 on the “t” axis (which way do we move along the “t”
axis?).
We know the answer (see above): t = ~1.0. This means that our “t” practically coincides with our “projected” opening
chess position on this axis!
What am I driving at?
At the fact that the psychological state of the great majority of practical players (or maybe all of them), before making
the first move of the game, cannot help but align itself with the psychological state of a chessplayer located within the
“hundred percent” Capablanca Algorithm! That’s what!
Your psychological state preceding the first move of a tournament game is one which belongs to you alone. It’s unique.
It is real, but it cannot be expressed in words. It is your baseline psychological state. It is the one we engage before
making our first move in the position after 17...Rac8.
And now for the second point: we will make a very important generalization, once again, in the game position following
17...Rac8.
Here, Capablanca first takes one step back (18.Nf1) before taking two steps forward (19.Ne3-d5).
What’s this – a strategic defense (18.Nf1) or a strategic advance (19.Ne3-d5)?
The answer is: both at the same time! More accurately, in the “100% Capablanca” the strategic attack is
indistinguishable from the strategic defense. The border vanishes; attack flows into defense. This becomes a two-sided
whole, an unbreakable bond which we shall henceforth refer to very simply as, strategic play!
The “formula” for strategic play is this: strategic play = attack/defense of vacant squares “as per Capablanca” – that is,
attack and defense of squares whose strategic value is less than that of pawns.
3) complete equality in the third factor of the position (there is an endgame on the board, with nothing threatening either
king).
Our preliminary diagnosis: the Tal Algorithm, with clear signs of the Capablanca Algorithm.
The algorithm in which “Tal” and “Capablanca” are present together is called the TC Algorithm. We’ll learn more about
it in Chapter 5. Suffice it for now to know that this is not some major new algorithm, but simply a hybrid of two
fundamental algorithms. The TC Algorithm has its own space, located in the neighborhood of a point on the “t” axis
with coordinates t = tcr = 1.25.
If the study position shows even minimal signs of a strategic algorithm, then we will always be aided by the fourth and
fifth factors of the position. They will make their contribution and provide a slight correction to our preliminary
diagnosis.
Our ∆k > 0: 8/28 > 8/32 (Black’s king and seven pawns stand within the b8-b5-h5-h8 rectangle, while the white king
and seven pawns are found inside the larger a1-a4-h4-h1 rectangle). The parameter ∆k impels us to the right, to the area
of genuinely Tal-like values of the “t” parameter (see the Algorithm Drift Chart).
Our ∆(10.Bd2) = 28/13 – 26/13 = 2.13 = ~0.15 > 0. Consequently, the fifth parameter of this position drives us, not to
the right, but to the left, toward the Capablanca Safety Zone.
The final diagnosis: both “Tal” and “Capablanca” simultaneously – in complete agreement with the preceding
diagnosis. Why?
Because in this position the fourth and fifth factors cancel each other out.
The game continued:
10...Bd7
11.Be2
11...e6 12.0-0
What for? The king’s place is not on g1, but on e2 – 12.Bb5 … Ke2, Rhc1...
14...Rhc8 15.a3
Better was 15.Nd2 right away, and if 15...f5 then 16.f4 … Nf3-e5, with chances to equalize.
Another tiny inaccuracy. The immediate 17.f4 looks stronger, when if 17...Rc7 then 18.Bxa5 Rxa5 19.Rxc7 Bxc7
20.Nb3 … Be2-d3, Kf2-d2, and White has every prospect of drawing.
17...b5
Too straightforward. 17...g5, delaying the b-pawn push until later, was more flexible. Black holds the strategic initiative,
but does it amount to anything real?
18.f3
Consistent (White strives for e3-e4), and bad! Janowski misses the chance to erect a fortress after 18.Bxa5 Rxc1+
(18....Rxa5 19.Rxc8 Bxc8 20.Nb3²) 19.Rxc1 Rxa5 20.Nb3 Ra8 21.Nc5 Rc8 (… 22...b6 23.Nd3 Rxc1+ 24.Nxc1 b4³;
21...Bc6 22.Nd3=) 22.b4. Now...
18...Nc4
19.Bxc4
The knight capture 19.Nxc4 is hardly better. After 19...bxc4 20.Re1 (… e3-e4) 20...Ba4 21.e4 Kf7 … ...Bb3, b6-b5,
...Ra4, b5-b4, Black holds all the cards.
The best move may have been 19.Kf2.
By way of exception, let us examine the diagram position from the opponent’s viewpoint – that is, from White’s side.
This will help us to subject Janowski’s move 21 to strong (but fair) criticism.
What does White have? His m = 1, t = 27/45 = 0.60 < tcr = 0.80. Complete parity in the safety factor, ∆k < 0, and
∆(20...Kf7) = 29/12 – 31/12 = -2/12 = ~-0.17.
The final conclusion: in the diagram position, White must play “like Petrosian.” He must defend. Why?
Because Black must attack, as he is not in the Petrosian Zone but in the Tal Zone. His “t” parameter is equal to t = 1/t =
1/0.60 = ~1.67. “Petrosian” is the antithesis of “Tal”!
And if all that is true, then the game move –
21.e5
– is nothing but a gross strategic blunder, since White has no right to attack. None! By elevating the e-pawn, Janowski
has bound himself forever to ∆k < 0. This means that now Black is assured play in the “gaps” of White’s position.
White cannot hold the game.
He had to trade on d5 with 21.exd5. After 21...exd5 22.f4 … Nf3 and Kf2, White’s ∆k > 0. The greater the ∆k, the safer
and more secure the position. After 21.exd5, Janowski would have had real drawing chances.
But after 21.e5, White no longer has any hope of saving the draw.
21...Be7 22.f4 b5
By combining play on the queenside (...b5-b4) with kingside action (...g7-g5, ...g5xf4, and placing the rooks on the g-
file), Black won on move 46.
Let me note that, in commenting on this game, I have more than once made use of earlier published commentaries. I am
speaking of the book edited by IM S. Solovyov.
In the afterword to this instructive fragment from Capablanca – Janowski, I want to say something important: we have
crossed the demarcation line that sets off the strategic attack on vacant squares from the strategic defense of vacant
squares.
More concretely: with 10...Bd7, Capablanca gave up control of the squares b1, c2, d3, and e4. He preferred the vacant
square b5 to these – a square in his own territory. In other words, Capablanca opted for defense over attack. Why?
Because the value of b5 was at least equal to the combined value of those four vacant squares in enemy territory!
My answer to that question was abstract. It is removed from concrete chess practice. So, what about practice? What
does a strong player think about when faced with a position requiring a strategic solution?
During the game, a strong player mentally reformats the position on the board into a more esthetically attractive one. He
paints a new picture in his mind. And then he turns it into reality.
In Capablanca’s head there spontaneously arise countless images. They flow together energetically. The final picture
(after 18...Nc4) doesn’t win right away, but it does win! It removes everything nonessential, everything twisted and less
than valuable – and therefore it’s truly beautiful!
14...g6
An error! More than likely, Capablanca did not see 14...Nxd4 15.cxd4 Be2!:
After the exchange, Point 1 of the strategic algorithm: Capablanca transfers his light-squared bishop to its best post.
Or 18.Bd3 Qc6 19.e5 Bxd4+ (19....Nd5 20.Be3 Qc7∞) 20.cxd4 Ne4!? 21.Bxe4 Qxe4 22.Qxe4 Bxe4 23.Rc3 f6∞
(Solovyov).
18...Qc6 19.e5
Very dangerous play! For White, that is, not Black. By “improving” the e-pawn – the advance guard of his position –
White condemns himself to permanent suffering along the a8-h1 diagonal. 19.Be3 is stronger, as this move corresponds
better to the Tal Algorithm (having played 18.Rh3 instead of 18.Bd3, White has set course for “Tal” and not for
“Capablanca”). The bishop move attacks the queen, throws a previously inactive piece into the cauldron of attack, and
offers a pawn!
A sample variation: 19.Be3 Bxd4 20.Bxd4 Nxe4 21.Qg4 f5 22.Qh4 Rf7 23.Bxe4 Qxe4 24.Rg3 Kf8. Fritz even rates
this position in White’s favor. Is it right? See for yourself!
19...Nd5 20.Qf2
Played for attack on the black king. White wishes to deliver mate on h7 – a dangerous, fateful delusion!
It was necessary to play instead 20.Bd3 … Bd2, with more modest intentions.
20...Bxd4 21.cxd4
21...Rac8 22.Bd1
22...f6
Simple and very strong. Black not only defends (f7 is freed for the rook), he also attacks (the threat is 23...fxe5 24.dxe5
Nxf4).
23.Qh4
An amazing disregard for self-preservation! 23.Bd2 … Rc1 was necessary in order to rid himself, at whatever cost, of the
pressure along the a8-h1 diagonal. After that, there could follow 23...fxe5 24.Rc1 Qxc1 25.Bxc1 Rxc1 26.Qd2 Ra1
27.fxe5 Nf4, with advantage for Black (Solovyov).
Besides 23.Bd2, the alternative 23.Rf3, going into deep defense, merits attention. But here too, obviously, Black is for
choice.
23...Rf7
24.Bf3
Some saving chances were offered by 24.Rf3 (Rybka). But there too, after 24...b5 … 25...Qb6 White’s situation is very
grim.
24...Qc4
An open attack on the d4-pawn.
25.Be3
Or 25.Qf2 Qxc1+-+.
Already an open attack on the queen. The attack is in full swing! An attack on lawful chess targets. The “Tal” of Black
is retribution for White’s thoughtless play.
27.Qe1
34...Kf7
Very elegant, and completely solid. By bringing up his king, Black sharply increases the packing of his king and pawns,
from 6/32 to 6/24.
The parameters of the final position: m > 1, t = 42/36 = ~1.17, “roughly =”, ∆k = 6/24 – 4/16 = 0, ∆(34...Kf7) ~3.00 –
2.14 = ~0.86.
White resigned. 0-1
At some imperceptible point, we crossed the line between “Tal” and “Capablanca.” That was at the point on the “t” axis
with the coordinate tcr = 1.25. Remember that we saw no armed border guards, only a solitary marker bearing a sign
with the inscription, “1 unit” (the chess value of a pawn).
We understood that “Tal” and “Capablanca” are not enemies, but good neighbors.
Now, in this game, we have begun our journey along the “t” axis somewhere near the border between “Capablanca” and
“Petrosian.” What awaits us?
We are awaited by a border marker and sign with an inscription: “1 unit.” No border guards there, either!
The final example in this chapter will focus on the theme of strategic defense – that is, on the theme of the Petrosian
Algorithm.
In the following position, Black stands much worse in the time factor. On the other hand, he is much better in the
compactness factor.
We need the fourth and fifth factors of the position! Our ∆k > 0 and ∆(13.e4) = 26/13 – 31/13 = -5/13 = ~-0.38. Which
means that both the compactness factor and the expansion factor push us in the direction of the Tal Algorithm.
Our final verdict is: both the “Capablanca” and the “Petrosian” at the same time.
The fourth and fifth factors made us drift away to a point on the “t” axis with the coordinate tcr = 0.80. We are
somewhere beyond it, on the border between the two algorithms...
13...Rd8
A natural reaction, given that exchanges favor Black. They are doubly favorable: because of compactness, and because
of “Petrosian.”
Curiously Black apparently has another solution besides the text: 13...Bb7 (pointed out by Solovyov). Play might
continue 14.e5 Rfc8 15.exf6 Bxf6. I ask you to extend it. At first without computer assistance!
Vidmar attacks – the white pawn crosses the demarcation line. Open attack on the knight!
White’s attack is quite legitimate: before the pawn shot, his t = 54/40 = 1.35 > tcr = 1.25. Confirm this!
15...Ne8
Covering the c7 and d6 squares. However, other, more aggressive, moves by the knight were also possible.
16.Bd3
16...Rac8
17.Nb5 Rd5
18.Nd6
The pinnacle of the attack.
20.Bxd6 Rd8
Winning the piece back and restoring material equality. The players soon agreed to a draw. A pretty finish!
Earlier, we summarized the preliminary results of our first acquaintance with the Capablanca Algorithm; it took place in
the no-man’s land between Games 15 and 16. There, we also outlined the most important instructions for studying this
algorithm.
I had promised you that:
1) we would go smoothly from the right-hand (advancing) part of the Safety Zone of the Capablanca to its left-hand
(defensive) side;
2) we would concentrate on the interaction between the fourth and fifth factors.
Now, in completing this chapter, we may state with satisfaction that we have both “gone smoothly” and “concentrated.”
Our forward advance was real, and we have evident successes in both the first and second directions of “advance.”
The transition from strategic attack to strategic defense has occurred almost painlessly. We crossed the boundary
between attack and defense at the point where t = 1.00, the point of symmetry. It cuts the Capablanca Safety Zone in
two. It also cuts into two parts the entire spectrum of chess attacks and defenses. This is a special point...
Attack and defense of vacant squares are tightly interwoven. They are two sides of the same phenomenon, two parts of
an organic whole which we will label strategic play, or play in the style of Capablanca...
From here on, the final formulation of the Capablanca Algorithm will look like this:
1) optimal placement of our pieces on squares suitable for a strategic attack or defense with respect to vacant squares;
2) advancing the pawns;
3) exchange.
This formulation is final, and it brooks no appeal!
A second direction is provided by the fourth and fifth factors of the chess position – that is, the factors of compactness
and space expansion. It is important to us that these two factors have little power, little intensity. Their intensity is
clearly less than that of the third factor.
The fourth factor is about the safety of the position. A compact position is a nearly impregnable one.
The fifth factor has to do with the tendency of a chess system (or chess position) toward limitless outward expansion. As
we “elevate” our position (the goal being to reach a queening square), we inescapably come into contact with our
opponent.
Expansion is effective only when
∆k ≥ 0. I repeat: only if ∆k ≥ 0!
The simplest way to increase your ∆k is by improving the king. Also by improving the pawns from the rear areas of the
position. In this manner the pawns are “enhanced” only to the level of the vanguard pawns. A fantastically powerful
method of strategic play! Why?
Because with this, we kill two birds with one stone. We not only increase our ∆k, but also our ∆(move).
We start our “advance” on the Petrosian Algorithm fully armed, as we have already acquired the attacking methods
known to all in the unified theory of chess play. These methods are the dynamic (“Tal”) and the strategic (the right-
hand “Capablanca”). Moreover, already we have not only acquired the left-hand “Capablanca,” but we have also even
succeeded in occupying a starting point for the Petrosian Zone. And all this, taken together, is our main achievement.
As we approach the Petrosian Zone, we need to understand well one simple thing: we stand “on the other side” of the
attack. For us, this means that the psychology of the chess struggle itself unquestionably must change: now we are
defending, not attacking.
Are you prepared, dear reader, to take the next important step?
The answer is obvious. Forward!
Chapter 4
3) a shaky parity in the safety factor, since the current happiness of both kings is temporary and insecure. The white a-
and b-pawns are prepared to break up Black’s queenside citadel at any time; on the other hand, Black has real chances
for a successful assault down the g-file.
So is our preliminary diagnosis “Petrosian”? Or is it still “Capablanca”? Or perhaps even “Tal”?
We have very serious problems, and alas, even the fourth and fifth factors are hardly in a position to help – not least
because these factors lack force and intensity. There’s also the simple reason that, here, the value of the parameters ∆k
and ∆(move) is close to nil. Verify this for yourself.
What to do?
The answer, dear reader, you will not get fully here or now. And not even in this chapter, but in the next one, where we
investigate the mixed algorithms – the TC, CP, and TCP algorithms.
Now for the diagnosis, without commentary. It’s a complex matter!
Our final diagnosis: the diagram position requires the TCP Algorithm. In other words, here it is possible for “Tal,”
“Capablanca,” and “Petrosian” to all show up together. For the chessplayer, the possibility of choice makes its grand
entrance!
Petrosian’s selection is one of the candidates, and it is genuinely surprising. There followed
17...c4
Petrosian: “After the game, I learned that this move surprised the onlookers. Indeed, its shortcomings are plain: the
square d4 becomes the property of White’s pieces. But only in theory, I say, because White can’t make anything of the
fact that the queen (or possibly the rook) will stand on d4. Whereas the knight, placed most favorably on this kind of
blockading square, is here deprived of the possibility of arriving at d4, as it is preoccupied with defending the e5-pawn.
Thus, Black’s hands are free to operate along the g-file. As we look at the rest of the game, we should not forget that
the threat of moving the bishop to g7 constantly hangs in the air, winning what is no longer the pride of White’s
position, but its weakness – the e5-pawn.”
I would add that the point of Black’s idea lies not in this aggressive and “anti-positional” move, but in the following
modest move by the a-pawn...
In the game, Spassky pulled the bishop back:
18.Be2
A mistake! 18.Bf5! (Petrosian’s mark) was stronger. The idea is to post the bishop on h3 and thus defend the g2-pawn.
After 18...exf5 19.e6 Bd6 20.Bxd6 Qxd6 21.exd7+ Rxd7 22.Nd4, White has some compensation for his two sacrificed
pawns.
Petrosian responded with
18...a6
Wonderful!
Now Petrosian’s plan becomes clear to us mere mortals: 19.a5 meets with 19...b5, and 19.b5 with 19...a5. In either case,
Black’s queenside suddenly becomes totally impregnable! Which means that after 18...a6 White is powerless to do
anything on the queenside.
Strictly speaking, we could close the curtain here – the theme of this chapter (the theme of defense) is exhausted for this
game. But we won’t rush; there’s still something to be learned from Petrosian: Petrosian the attacker, Petrosian the
virtuoso of strategic play.
19.Kh1
What does Black have? m > 1, t = 33/35 = ~0.94, a small “+” in the safety factor, ∆k > 0, and ∆(19.Kh1) = ~0.10.
Diagnosis: “Capablanca”? TC Algorithm?
Note how the position has changed radically in just two moves. Black’s progress in the second, third, fourth, and fifth
factors is obvious (in the position after 17.a4, ∆k = ~0.00 and ∆(17.a4) = ~0.11).
Need we comment on these irrefutable facts? I don’t think so...
After Spassky’s move, 19.Kh1, Petrosian begins to play, not in “Petrosian” style, but in “Capablanca” style. The
position requires it, and Petrosian complies quietly with these requirements.
What is the core requirement for a strategic attack in the Capablanca Algorithm? We know the answer: when attacking,
the stronger side must increase the ∆(move) of his position. With ∆k ≥ 0!
Why does “Capablanca” require ∆k ≥ 0 (let me add that “Petrosian” accepts this requirement with delight)?
Answer: because, when ∆k ≥ 0, our opponent’s counterplay is practically nil. It is smothered at birth – prophylaxis!
What do we have now? m > 1, t = 42/30 = 1.40, a small “+” in the third factor of the position, ∆k > 0, and ∆(23.Rad1) =
~0. 29. We note that ∆(23.Rad1) > ∆(19.Kh1) = ~0.10.
Our diagnosis after 23.Rad1: “Tal”!
What does Petrosian do?
He doesn’t shrink away from sacrifices! There followed:
The variation 26...dxe4 27.Bxe5 Qxe5 28.Qd8# would, of course, not be to Black’s liking.
30...f5
31.exd5 f4 32.Qe4
Or 32.Qa7 g3 (Kasparov).
Or 39.Bxh3 Qd7-+.
41...Ng4
42.fxg4 f3
Petrosian continues to increase the ∆(move) of his already “elevated” position. Incidentally, he also threatens mate on
h1.
“Capablanca” at the service of “Tal”? Or is it “Tal” at the service of “Capablanca”? No need to answer – these are
rhetorical questions...
43.Rg2 fxg2+
31.R1c2
To what end?
In order to march his king from f1 to b1! Petrosian is in no hurry...
Black’s situation is grim: he has no counterplay to speak of and stands half-paralyzed. He must sit and wait.
I say at once: I have serious doubts about Black’s play. The g8 and h7 squares are not the best ones for the king. The
best square for it is g7. Why?
Because on g7 the king stands more compactly!
One more thing: the bishop may stand better on e7, closer to the king, than on d6.
36.Qe2
Preparing g3-g4.
36...Qb7 37.Rc1
The immediate 37.g4 looks stronger. One possible variation, according to Rybka, goes 37...hxg4 38.Qxg4 Qe7 (38...Ra6
39.Rxa6 Qxa6 40.h5±) 39.h5 g5 40.f4 Rg8 41.f5 e5 42.Qg2 e4 43.Rb6±. The threat is 44.Rb5 – the target on d5 is
added to the one on a5. Plus, there are other threats...
37...Kg7
Unfortunately for Unzicker, Black can’t play 37...f5. Why? Because, after invading on b5, followed by Rc6-b6 and Rc1-
c6, Black’s e-pawn drops.
38.Qb5
Petrosian would not be Petrosian if he were in a rush to strike. Petrosian doesn’t rush – why?
Because the situation is psychologically very difficult for the opponent, who is relegated to conducting a deep and
hopeless defense. The probability that he will make a mistake grows with every move.
38...Qa8
38...Qxb5 also loses. After 39.axb5 a4 40.b6 Rad7 41.Na5 Ra8 42.Rxd6 Rxd6 43.b7 Rf8 44.Nc6 Rxc6 45.Rxc6 Rb8
46.Rb6 b3 47.Kc1 … Kd2-c3-b4, Black can resign with a clear conscience. Bravo, Rybka!
39.f4
A debatable decision. Most likely, Petrosian played this so as to prevent a probable ...g6-g5 in reply to h4-h5 – see the
note to 37.Rc1.
The downside of this move is that now yawning cracks appear in White’s kingside...
39...Kh7
Wrong! This is a fatal error. The king stood ideally on g7, and there was no need for it to move from there.
Rybka recommends 39...Rb8, when if 40.Qe2 (… 41.g4), then 40...Be7 and it turns out that 41.g4 is bad, since after
41...hxg4 42.Qxg4 Rh8 White’s attack dries up – and Black seizes the initiative!
Therefore Petrosian would have had to refrain from g3-g4 and play e3-e4 instead, and perhaps even seek to bring the
knight to e5 without hurting himself – he must not allow the ...b4-b3 wedge!
Everywhere White is better, but does he have real winning chances?
Whereas now...
40.Qe2
43...Qf6 44.Ka2
44...Kg7
Too late!
46...Be7 47.Qf2
47...Kf8 48.Nd2
Scaring his opponent with shadows. White wants to bring the knight to e5?! Yes or no?
48...Rb7 49.Nb3
49...Ra7 50.Qh2
51...Rad7
Your task is to find the win after 51...Rxc8 52.Rxc8+ Ke7. Without the aid of Fritz, Rybka, and their cohorts.
52.Nc5
Awesome!
52...b3+
A double blow!
54...Rb6+
55.Ka2
3) A minimal advantage in the third factor (White’s h5-pawn presses a little on Black’s castled position).
The immediate final verdict: the Tal Algorithm. Formally, our situation is simple, because White stands better in the
second and third factors. But... there is a subtle detail.
The subtlety is that, in our position, ∆(18...Be8) = 1.00. Check it out!
This value for ∆(move) is a great rarity in a middlegame with numerous pieces still on the board. Here we have thirty-
two of them...
19.h6
A non-standard decision! For Petrosian, the great significance of ∆(move) is more important than any attack on the king.
Truly, a bird in hand is worth more than two in the bush!
Most players would prefer 19.hxg6, followed by doubling rooks on the h-file, Be3-h6, trading dark-squared bishops,
and then attacking the enemy king relentlessly. “Tal” is “Tal,” and the king is always (always?) the primary target in
the attack.
19...Bh8
Petrosian’s essence is not an attack on the king, but defense. He defends based on accepted chess principles – here, the
enormous significance of ∆(move) in White’s position after 19...Bh8.
A most pleasant side effect of the perilous invasion of h6 is that Black’s dark-squared bishop is forced into a remote
corner. Now, after 19...Bh8, White in effect has an extra piece!
Returning to the game now. The standard question: what now?
The standard reply: always, everywhere, and under all circumstances we must act in total compliance with the
requirements of the position. Always and everywhere!
Question: what does the position require after 19...Bh8?
Answer: in a word, “Tal.” But this “Tal” will no longer be the “Tal” of the kingside attack. It will attack queenside
targets, where White has an extra piece.
20.Na4
In order, after 21.b4 cxb4, to take on b4 with the queen – followed by c4-c5 with a crush!
Now our m = 1, t = 41/20 = 2.05!, a small “+” in the third factor of the position, ∆k < 0, and ∆(22...Nd7) = ~1.33!
Our progress is obvious: in the space of four moves, White has added to the second and fifth factors. The price of
progress is a “permanent” ∆k < 0. And this “poor” ∆k will now become a constant headache. This parameter will
always keep us in shape – we cannot flag in our efforts...
23.b5
The opponent’s position along the c-file cannot be cracked. So we open the next file!
Exchanges are a double-edged sword for Black. On the one hand, exchanges are favorable – Black has the more
compact position. On the other hand, White has an “extra” bishop...
Black’s position is strategically hopeless!
The parameters of this position are: m = 1, t = 39/26 = 1.50, a small “+”, ∆k < 0, and ∆(29...Nxb6) = ~0.92.
Our diagnosis: the Tal Algorithm.
30.Qb2
If 32...Ndb6, then 33.Be2 … 34.Bd1 and... exchanges, exchanges, exchanges. The horizon of White’s dreams, hardly
fantastic ones, is a four-piece ending. Black has a king and the bishop on h8, while White has... need I continue?
Trades are also a double-edged sword for White (see the note to 26...b6). On the one hand, they favor Black, since he
has the more compact position. On the other hand, White does have that extra bishop...
White’s position is strategically winning!
Intending the following setup: Ng3-e2-c2 with Bd3 and Bd2 (if the black queen goes from a8 to a3), and then Qc2-a2
with a queen check that would be very unpleasant for Black.
Attention! The bishop is needed on d3 in order to hold the f5 square one more time. We cannot relax – we must always
bear in mind the possibility of ...f6-f5. On f5 (after a sacrifice) there must not appear an awkward and helpless white
pawn, but rather a proud piece – be it the queen at c2, the d3-bishop, or the g3-knight!
Thank you!
36.Qxa4 Nxa4
A minor-piece ending!
We have: m = 1, t = 20/13 = ~1.54, rough equality in the safety factor (attacking the king in the endgame is a fantasy),
∆k << 0 (danger!), and ∆(36...Nxa4) = 0.50. Calculate these numbers for yourself.
Looking at the position, I am convinced of a sad fact: we have gone down. More exactly: our position has deteriorated
seriously in the fourth and fifth factors – compare the parameters of this position with the corresponding parameters of
the previous positions. Moreover, we have allowed the enemy to invade our territory – we are under attack! The threat
is 37...Ndb6 and 38...Nb2, winning a pawn.
And nevertheless, we are... two steps (two exchanges) away from victory! An inevitable and crushing victory. Why?
Answer: because there are so few pieces left on the board!
We must take a minimal amount of prophylactic measures, and Black’s attack (or “attack”) will peter out.
37.Ng3
White not only puts the kibosh on a possible ...f6-f5, but also holds the c4-pawn.
Incidentally, all of our fears concerning ...f6-f5 were clearly overblown, since the f-pawn’s advance would meet with
g4-g5, leaving the h8-bishop in an unenviable position. A sad situation!
War is hell...
37...Ndb6
38.Bc1
Necessary, and sufficient to beat back the attack irretrievably. Black has no forces left with which to continue the
struggle... the bishop on h8... hopeless despair... the beginning of the end?
Reader! I call upon you, I require of you: broaden your horizons! Expand the “Tal” with the “Petrosian,” and let
“Capablanca” help you! Believe and know: chess is One...
There followed:
38...Nc3
39.Ke1 Ke7
Alas! If 39...Na2 40.Bd2 Nb4 (… 41...Na4), then the simplest win is 41.Kd1 Na4 42.Kc1 followed by Bf1-e2, Ng3-f1-
e3, Kc1-b1 (useful prophylaxis against a possible check at d3), and Be2-d1. And then the “impregnable” Black fortress
falls to pieces!
40.Kd2
40...Nca4 41.Kc2
41...Nc8
42.Kb3 Nab6
43.Bh3
43...Kd7 44.Bd2
From this square the bishop eyes, not only a5, but also g5!
44...Ne7
45.g5+
In the position just before this check (see diagram), White has: m = 1, t = 22/17 = ~1.29, roughly “=”, ∆k < 0, and
∆(44...Ne7) = 1.00. Our diagnosis: the TC Algorithm – that is, “Tal” and “Capablanca” at the same time.
As computer analysis shows, Petrosian’s move is correct. With this move, White forces the win, and there can be no
doubt that Petrosian calculated everything to the end. The final position at the moment of Black’s capitulation was not a
pleasant surprise for him: he foresaw it, more or less accurately...
Why then, you ask, does Petrosian’s radical and “anti-positional” decision (that bishop on h8!) have a right to exist?
I will answer that without commentary, for now (that’ll come in the afterword to this game): “Tal” was more important!
More important than all our blessed undertakings connected with the bishop isolated on h8. “Tal” is not forbidden!
45...Ke8 46.Ba5
An open attack.
46...Nec8 47.Nh1
Clearing h3 for the knight and preventing the bishop’s activation with ...g5-g4.
49...Bd8
If 49...Nd7 (Fritz 11’s recommendation), then after 50.Ka4 Bd8 51.Nh3 Ne7 (...Ne7-g8xh6 – that’s why Black allowed
the white king into his camp without a fight) 52.Nxg5 Ng8 53.Nxh7 Nxh6 54.Be6, White has all the winning chances.
Your task is to find the win. Here’s a hint: the value of the empty square b6 is, for the king, not less than the value of the
pawn at g6, to put it mildly. Enjoy!
As we have seen, with the text move (49...Bd8), Bertók chose the path of war over the path of non-violent resistance...
55...Nxc8 56.Bg5
At last!
61...Nb6 62.Bd8+
To an outside observer (that’s you and me), Petrosian was a player who preferred the Capablanca
Algorithm to the Tal Algorithm, and the Petrosian Algorithm to that of Capablanca.
17...Bxc3
Remarkable! Why?
First and foremost, because now Black is freed forever from some very serious unpleasantness in the case of b2-b4. But
that’s not all. The main thing is that Black has traded off a useless bishop for White’s very good knight! The bishop is
useless because it fires only at the dark squares, whereas the g4 square is a light one. Petrosian feared a possible break
along the g-file, and not without reason. Almost certainly, he was right!
The uncompromising Fritz requires 17...b5, with 18.cxb5 Nb6 etc., to follow. Rybka treats the position more carefully:
17...Kh7, 17...Qe7... Petrosian won’t go that way!
18.bxc3 Nf6
White has two pieces controlling g4, Black has three. And that’s not all.
19.a4 Kh8
“The following moves by Black are aimed at preventing the dangerous break g2-g4. Petrosian comes up with the
following plan: to set up his pieces on the g-file and at the right time play ...h6-h5. Black has enough space to carry out
the necessary regrouping on the kingside. The important thing is that he does not have to worry about queenside threats,
where the possible threat of a breakthrough has been liquidated” (Sokolsky).
20.Nf2 Rg8
5 > 4.
23.Qd2
White intends to advance his pawn to g3 and then double rooks on the g-file. Perhaps it was stronger to play 23.Rb1,
preventing the free advance of the rook from a8 to the “precious” g-file.
23...Bd7 24.g3
Too straightforward.
24...Rae8
27.Rg2 Reg7
7 > 5. Overprotection!
In the final position, m = 1, t = 24/31 = ~0.77, “rough equality”, ∆k = 0, and ∆(29...h5) = -0.29. These parameters are
close to those of the position after 17.Nh3.
Black is out of danger, as is White. Parity! The players agreed to a draw. ½-½
The following game between Petrosian and the master Schweber is in fact only an extension of the central theme of the
preceding game: hunting down the opponent’s good knights. For Bronstein, that was his c3-knight, while for Schweber
it became his f6-knight.
18.Bb5
And here’s the answer to that not especially difficult question: the knight White has liquidated can no longer come to its
best square – d6.
Prophylaxis above all, no?
20...Rxe8
After the exchange, as we stated in “Capablanca,” Petrosian embarks on a new setup for his pieces. The first point of the
strategic algorithm in action!
In order, on 23...Bxc4, to have the proper reply 24.Rxc4, when the f3-pawn is safe.
23...Bf8 24.Ng3
On to f5!
24...Bc8
“The poor bishop wanders without knowing what to protect first – c4 or f5” (Goldberg).
Of course! With this move Petrosian considerably increases the density of packing of his king and pawns.
26...Ra7 27.Rf2
On to c2!
29.Nxe5
The key to success is compactness around the king. The larger this factor, the easier it is for the weaker side
to hold the position and the more difficult it is for the opponent to get to his material targets.
25...Re6
Tal: “This purely positional sacrifice (a quiet move, with no checks or apparent threats!) had an indelible effect on me.”
Kasparov continued (quoting from My Great Predecessors, Part III): “... For the sake of what?! In order to block the
advance of the e-pawn, and also to open the way for the knight to d5.”
The dangers looming over Black were truly great. For example, here’s what would have happened were it White’s turn
in the diagram (variations by Rybka):
26.Bf3 Ra7 27.d5 Ne7 28.d6±;
26.Bf3 Na7 27.d5 Qd7 28.d6 Ree8 29.Bd5±.
Now let’s return to the position after 25...Re6. After the text move, it will be Black, not White, who enjoys hegemony in
the center!
26.a4
Probably the immediate 26.Bxe6 was stronger, and if 26...fxe6, then 27.Rg3! (Kasparov). Next White moves his bishop
to d2 and prepares to attack the king. White stands better.
Therefore, after 26.Bxe6, instead 26...Qxe6 is stronger (Kasparov). Kasparov continues: “After the possible 27.Rg3 Ne7
28.h4 Nd5 29.Qg5 Rd7 30.h5 h6 31.Qh4 Bd3 the problem of the b2-bishop cannot be resolved immediately, but White
retains some advantage...”
White might have greater prospects after 28.Rg3 b4 29.Qg5 (Rybka’s recommendation).
31...b4
“Reshevsky’s clever play combined with Petrosian’s iron logic make this game one of the tournament’s jewels... [N]ow
it is White who must solve a difficult psychological problem: should he exchange on b4, which practically guarantees
the draw, or advance the pawn, driving out the knight and obtaining winning chances, as well as losing ones?”
(Bronstein)
Curiously, here he could also have played the non-human move 31...bxa4 (Rybka). Check to see if this is the case.
One thing is clear: Petrosian succeeded in equalizing (or almost equalizing) the position. Credit for this goes to
25...Re6!.
32.cxb4 axb4 33.a5 Ra8 34.Ra1 Qc6 35.Bc1 Qc7 36.a6 Qb6 37.Bd2 b3 38.Qc4 h6 39.h3 b2 40.Rb1 Kh8 41.Be1
In the afterword, we will try to answer one question: just how paradoxical was Petrosian’s move 25?
25...Re6 enabled Black to make a draw. It could be that this was the strongest move. But... but chess is truly
inexhaustible!
Let me reformulate the question and put it to you this way: did Petrosian’s move fit the logic of our preceding
considerations? Yes or no?
After giving it some thought, we answer, “Yes,” because nobody can forbid us to sacrifice! When sacrificing material,
the weaker side doesn’t give it up in return for nothing – he does it in order to achieve some other (non-material)
benefit. By sacrificing rook for bishop, Petrosian gained control of d5. To him, the chess value of that empty square
was comparable to the value of the sacrificed material.
We don’t need to worship chess material. The material element in a position is not the position itself, but only one
aspect of it, a part of the whole. And this part is not capable of outstripping the whole...
“Tal” sacrifices in superior positions are offered for the purpose of increasing the tempo of the attack. This is a non-
equivalent exchange. The ultimate goal is to mate the enemy king.
“Petrosian” sacrifices material in inferior positions in order to slow down the attack. The goal here is to convert material
chess values into strategic ones. The ultimate goal is to defend the king.
3) approximate equality or even a slight inferiority in the safety factor (within the rectangle f1-f8-h8-h1, White has an
extra queen; on the other hand, the black king is more tightly “bundled”).
Our preliminary diagnosis (which, as we’ll soon explain, will become final): the Petrosian Algorithm. And our
“Petrosian” will above all be strategic, rather than dynamic. Why?
Because our king is tucked away very solidly!
The fourth parameter: ∆k = 0.
The fifth is: ∆(31.Qg4) = ~-0.55. This factor nudges us slightly to the right along the “t” axis – that is, moving from the
Petrosian Algorithm to the CP Algorithm. This small drift to the right, of course, cannot set us beyond the boundaries
of the Petrosian Zone. We have dug deeply into it – our starting point is over to the left of
tcr = 0.80.
31...Rf4
A sacrifice! To what end?
To free himself from bothersome strategic chains and to control the strategically important point e5.
What’s more important: a rook for a bishop, or an available center square?
A very difficult question. So difficult, in fact, that Tal himself did not find the answer!
32.Bxf4
“Of course, if Tal had understood all the way what taking the exchange would lead to, he would have contented himself
with a pawn: 32.Rxf4! exf4 33.Bxf4. I believed this situation was better for Black than playing on with material
equality but a very cramped position” (Petrosian).
It is quite possible that then (after 33.Bxf4) there would have followed 33...Qf6 34.Bxd6 Qxd6 35.Rf1 Ne5 36.Qf4 h6 –
Rybka’s variation. After this White would have m > 1, t = 32/28 = ~1.14, approximate equality in the safety factor.
Diagnosis: the Capablanca Algorithm. I add that White would have good winning chances...
But now the battle turns considerably sharper – now Black has counterplay!
32...exf4 33.Nd2
If 33.h6, then 33...g5 is very possible (and 33...g6 looks good, too), i.e. 34.Rae1 Ne5 35.Qf5 f6 36.Nd2 Nf7. Who’s
attacking whom?
On the other hand, Black’s position is nevertheless inferior: 37.Nf3 Nxh6 38.Qe6+ Kg7 39.e5² (Rybka).
33...Ne5
34.Qxf4
Wrong! Tal evidently was blinded by the beauty of his aggressive idea – see his next move. Either of two other queen
moves was stronger: 34.Qf5, or even 34.Qh3 (suggested by Kasparov).
34...Nxc4 35.e5
Question: Is Tal entitled to play like “Tal”?
35...Nxe5
36.Ne4 h6 37.Rae1
“After 37.Nxd6 Qxd6 Black would have supported his knight by ...f7-f6 and would have had an impregnable fortress
plus counterplay with ...c5-c4. But, of course, White should not have allowed that which occurred in the game. 37.b3!?
came into consideration”
(Kasparov).
37...Bb8
A powerful move! The silly little threat of 38.Nd6+- is warded off; and that’s not all...
38.Rd1 c4
A cold shower in retribution for unjustifiably raising his sword. Now it’s Black who’s on the march – he threatens
...Ba7+ and ...Nd3. Unexpectedly for him (not for us, though), Tal has become the weaker side...
Here I end my commentary: the theme of defense in this game has been exhausted. The game, not without mutual
errors, eventually ended in a draw on move 73.
The Petrosian Algorithm is the contrary of the Tal Algorithm. Later on we will defend difficult positions, teetering on
the edge between draw and defeat. But we will not engage in post-mortems, the theme of defending hopeless positions
– this is not our focus.
If we are the weaker side and our opponent attacks us for legitimate chess reasons, then we are obligated to humbly
surrender to him what he deserves for being the stronger side. Without a struggle and almost with a feeling of deep
gratitude, we surrender sacred space to our opponent. This way we concentrate our forces. Compactness is our aim!
If the opponent is very strong and if he has enough forces to achieve a material advantage, we may have to clench our
teeth and consent to this terrible situation.
What do we have to defend at all costs? You know the answer: the king! Because the king is a piece of limitless value...
9...Nf7
He retreats. Of his own accord! Pressing on the spring, Black yields to his opponent some of his space. In other words,
Petrosian increases the localized density of his own pieces right around the black king. The king is at the center of this
little village!
10.Qd2
Petrosian: “It was probably time to play d4-d5, which would have made the drawbacks of Black’s position stand out
more clearly.”
I suspect that Petrosian, if he had White here, would have played 10.d5, and if 10...Ne5, then 11.Nd2 – White doesn’t
need trades! The check 11...Nd3+ is not a problem – see for yourself!
10...c5 11.dxc5
“And here 11.d5 was possible. White plays very consequentially. Having an edge in development, he attempts to open
up the game, figuring to make use of the better mobilization of his pieces.
“From the viewpoint of chess commandments, it all looks quite reasonable, but in Black’s camp there are no vulnerable
points...” (Petrosian).
Petrosian continues: “Castling seems more natural, in order to bring at least one rook into play and put the king away in
a safer place. But after this it would be difficult to get in ...Nb8-c6. The natural knight sortie to d5 (with tempo!) would
gain in strength, and if the king remains in the center, he simultaneously protects the pawn at e7 and the bishop at d7.”
I would add: the bold sacrifice of the c5-pawn also has great psychological significance: it is difficult to make the
mental shift from attack to defense. It might seem that after 13.Bxc5 Na5 (or 13...Nce5), Black regains the initiative.
Here’s a question for White: what is he to do?
I reply: He must jump, without looking back, into the complications Black has provoked, since the pawn sacrifice has
no “right” to be 100% correct – Black is worse in the “t” parameter, and therefore there is no justification for playing
“like Tal.” By sacrificing this pawn, Black hopes only to get his material back. He wants to avoid a catastrophe in the
“t” parameter, at least for a while. At the cost of a pawn, in order to save the king from the looming threat of assault!
13.Kb1
13...b6 14.g4
In order to defend the target on d7 again (certainly not overprotecting!). Rybka’s recommendation of 14...Qb7 …
15...Rd8 doesn’t look bad, either.
15.Rhe1 Rb7
16.e5
Rybka recommends 17...Nxg5 18.Bxg5 Be6 19.Nd5 0-0 with approximate equality.
18.Nd5
Petrosian gives this move a question mark. Moreover (he writes), “I would have played 18.Bd5 e6 19.Nxf7 Rxf7
20.Be4, with a further advance of the h-pawn and the threat of Nc3-b5-d6.”
Very strong! Although Petrosian considers that he could also defend after 20...Bxh6 21.Qxh6 Rxf2 22.Rf1 Rf7. See for
yourself! Use a computer for assistance.
The decisive error. However, to give him his due, we note that after 24.Rf1 Bc6 (or 24.f3 Bc6), Black retains his
positional edge.
24...Qc7
25.Rg5
25...Rf4
25...Qh2 (Rybka) is also very strong. After 26.f3 Bc6, White is helpless.
26.Qd3
It is simpler to resign.
27.a4
a) 27.Nb5 is the most aggressive choice. White hits the target on a7 – being worse, he plays for an exchange. The fewer
targets, the easier the defense!
At first, I admit, the knight’s leap out to b5 looks crazy: 27...Bxa2 28.Nxa7 Ra8. After 29.Ra1 Bd5, White stands clearly
worse: see for yourself! But then 29.Nc6 (bravo, Rybka!) – draw!
However, Black has better moves than the primitive 27...Bxa2. One such is 27...Rf4. Now 28.Nxa7 is bad, because after
28...Ra4 29.Nc8 (the only move) 29...Nxc8 30.Bxc8 Rxa2, Black is almost winning. So taking on a7 is out. Much
better is 28.Nc3, when it looks like White has every hope of holding the position: 28...Bc6 29.g3 Rd4 30.Re2.
White also has excellent drawing chances after 27...Rb8. Both 28.a4 and 28.Nc3 are possible. Sample variations:
a1) 28.a4 Rb6 29.Nc7 Bxg2 (Rybka is amazing) 30.Kxg2 Rc6 31.Nxe6 Rxa6 32.Nc5 Rc6, and now either 33.Nb7 or
33.Nd3 should lead to a draw;
a2) 28.Nc3 Rb2 29.Nxd5 exd5 30.Rf1 Rxa2 31.Bb5 Rb2 32.Bd7 (… e5-e6, Rf1-f7 etc. with strong counterplay for the
sacrificed pawn) 32...Re2 33.e6 Ng6:
27...Rf4 28.Rd1 Ng6 29.Bc8 Kf7 30.a5 Nxe5 31.a6 Rg4 32.Rd2 Nc4 33.Rf2+ Ke7 34.Nb5 Nd6 35.Nxd6 Kxd6
36.Bb7 Bxb7 37.axb7 Kc7 38.h3 Rg5 39.Rb2 Kb8 40.Kf2 Rd5 41.Ke3 Rd7 42.Ke4 Rxb7 43.Rf2 a5 0-1
30...b5
We can only marvel at Petrosian’s courage! And then we ask: hasn’t the Armenian GM overestimated the defensive
resources of his position?
I answer: more “yes” than “no.” Here’s why:
Contrary to Kasparov, after 30...Bd6 Rybka confidently holds the position! For example, 31.a5 Rc7 (safely protecting
b7) 32.Nb6 Qe8 (taking on b6 is dangerous – check it out!) 33.Rcb2 Qf7.
What next? I don’t see anything. In the search for a decisive blow, Rybka bobs and weaves: 34.Bf3 Rd8 35.Rb1 Be7
36.R3b2 Bd6 37.Be2.
Look at this! Do you see the shots on d5, b7, and a6?
37...Nb8!. Everything’s in order!
Conclusion: after 30...Bd6, Black need not lose – while, after the text move (30...b5), Black’s position falls apart.
31.axb5 cxb5
If 31...axb5, then more than likely there would follow 32.Ra2+ Kb7 33.Qb1, when Rybka continues 33...Re8 34.Na5+
Bxa5 35.Bxa5 Kc8 36.Bb4 with decisive threats. Thus, for example, 36...Nxb4 37.Rxb4 Rb6 is bad as 38.Bxc6
decides. Nor does 36...Qd8 save him – after 37.Bd6, it looks like Black can’t avoid losing material. Verify this!
32.Ra2
A powerful move – the top line from the powerful chess engine. Now 32...bxc4 is bad, as 33.Rxa6+ mates. What to do?
The standard reply is: determine the parameters of the position, “calculate” the algorithm, and then find the strongest
move, if you can – that is, the one which satisfies the needs of the position.
And one more thing: be fearless. Like Petrosian! Like Kasparov! Like any great chessplayer...
Black to move – and I am sure, dear reader, you know what to do!
We have: m > 1, t = 41/42 = ~0.98, a very large “minus”, ∆k = 7/21= ~6/25 > 0, ∆(32.Ra2) = ~-0.46. The diagnosis:
“Petrosian,” “Petrosian” being concrete in nature: the black king is in great danger! Why?
Because the packing density of the black pieces immediately around their king has decreased significantly...
What to do, then?
Petrosian gives the answer, and it is a king move.
32...Kb7
Pursued by his enemies, the king runs for cover. In some variations (let’s say, after 33...Re8) he has chances to hide on
the c8 square or still further away on d8. Obviously, the fate of the king and the game hangs for the moment on the
thinnest of threads. We must calculate, calculate, and calculate again a multitude of variations. Playing by general
considerations takes a vacation for both White and Black!
Here are some variations, which I have borrowed from My Great Predecessors and fleshed out with Rybka’s help:
a) 33.Na3,
a1) 33...Bb6 34.Nc2 Ra8 35.Nb4 Qd6 36.Rb1 (if 36.e4, then 36...Nc5 – Timman’s move – 37.dxc5 Qxc5+ 38.Kh2
Qg1+ 39.Kg3 Qf2+ 40.Kh2 Qg1+ with perpetual check) 36...Nb8 37.Qb3 Rc4 38.Rc1 Rxc1+ 39.Bxc1, “with an
overwhelming advantage for White.”
a2) 33...Nb6 34.Nxb5 axb5 35.Qxb5 Rh8 36.Bb4 Qd8 37.Qa6+ Kc6 38.Bc5 Kd7 39.Qb5+ Kc8 40.Ra7, with
unstoppable threats, for example 40...Qe8 41.Qa6+ Kd8 42.Rxc7 Kxc7 43.Bxd5 exd5 44.Bxb6+ Kd7 45.Qa7+ Ke6
46.Qxg7+.
But!... But Black has a miracle move that Kasparov missed:
a3) 33...Qe8 – deep prophylaxis! And moreover – if we can believe Rybka – Black has every chance to draw. He is no
worse; a sample variation runs 34.Nxb5 axb5 35.Qxb5+ Bb6 36.Bxd5+ exd5 37.Qxd5+ Kc7=. Or if 34.Nc2 (…
35.Nb4), then 34...Bd6, and again White has no prospect of crashing past the opponent’s defenses. Approximate
equality! Or are you doubtful?...
b) 33.Bb4 – the game move, which we will investigate a little later.
c) 33.Qb1 – the last of three moves worth looking at, which is also in the first line by Rybka.
c1) 33...Rg8 34.Na5+ Bxa5 35.Rxb5+ Ka7 (35...axb5 36.Bxa5 [… 37.Qxb5++-] 36...b4 37.Bxb4 [37.Qd3±/+-] 37...Qd8
[37...Qf7 38.Ba5+ Kc6 39.Qc1+ Kd6 40.Qa3++-] 38.Be7+ Qb6 39.Rb2+-; 35...Bb6 36.Rxd5 exd5 37.Bxd5+ Kc7
38.Qc1+ Kd8 39.Bxg8+-) 36.Raxa5±;
c2) 33...Re8 34.Na5+ Bxa5 35.Bxa5 Qd6 (35...Ka7 … 36.Bf1 37.Bxb5+-) 36.Rxb5+ Ka7 (36...axb5 37.Qxb5+ Kc8
38.Rc2+ Nc7 39.Rxc7+ Qxc7 40.Qa6+ Kd8 41.Bxc7++-) 37.Bd2±.
As we can see, White enjoys a clear advantage in variations (c1) and (c2). He has an attack on the king, with material
equality to boot – and White has managed to eliminate that daring pawn at b5...
I repeat my question: when playing 30...b5, didn’t Petrosian overestimate the defensive resources of his position?
Kasparov attacked the queen:
33.Bb4
Petrosian retreated:
33...Qe8
“From here the queen indirectly defends the b5-pawn. 33...Qd8? was much weaker in view of 34.e4! fxe4 35.Qxe4 with
irresistible threats: 35...Qe8 36.Qxd5+! exd5 37.Bxd5+ Ka7 38.Rxa6+! Kxa6 39.Ra3+ Ba5 40.Rxa5#, or 35...bxc4
36.Qxe6 N7b6 37.Rb1 Qd7 38.Qxd7 Nxd7 39.Bxd5+ Ka7 40.Rba1 Rb6 41.Bxc5 and wins” (Kasparov).
I would add: along with 33...Qe8, 33...Qf7 also doesn’t look bad; Rybka recommends it. Black can be confident about
holding the position.
34.Bd6
If 34.Ba3, then at least 34...Ra8 … ...Ra7 and ...Ka8, tucking the king away safely, is possible.
Nor is the check on d6 dangerous: 34.Nd6+ (Petrosian’s recommendation); after 34...Bxd6 35.Bxd6 Ra8 36.e4 Rc1+
37.Kh2 fxe4 38.Bxe4 N7b6, Black is no worse (Rybka).
34...Ra8
35.Qb1
“White again plays on general grounds, hoping that an opportunity for him to land some combinative blow will present
itself. After the ʻconcrete’ 35.e4 fxe4 36.Qxe4 Qf7 Black could have felt quite calm with his powerful knight at d5. But
35.Rb1! Bb6! (35...Kc6 36.Ba3!, and if 36...Nb6, then 37.Na5+ Kd7 38.Rxb5!) 36.Ne5 would still have retained
adequate compensation for the pawn” (Kasparov).
While, after the text move, there came something unique, something extraordinary...
35...Kc6
Kasparov’s commentary on this move (and first of all from a psychological viewpoint) is interesting: “A fantastic
defense! This move, which Petrosian made instantly, threw me into complete confusion: how is it possible to move the
king with a board full of pieces? After Steinitz, who had done such a thing? The psychological effect of the 30...b5!?
thrust and the king march ...Ka7-b7-c6 was so strong, that I was unable to gather my thoughts and I quickly lost.”
36.Rba3
After this, White loses in all variations! He has to take the bishop: 36.Bxc7. Then it would be quite possible to reply
36...bxc4 37.Rb7 Rxc7 38.Rxa6+ Rxa6 39.Qb5+ Kd6 40.Qxa6+ Ke7 41.Bxd5 Rxb7 42.Bxb7 Qb8 43.Kf2 (following
Kasparov’s line of thinking) – since the king is inside the square of the c-pawn! The position would be approximately
equal, with the most likely outcome being a draw. But now...
36...bxc4
Now White has absolutely no compensation for the sacrificed piece.
When defending, we defend our material and non-material chess values against our opponent’s aggression. We defend
them according to our known value scale – from the king all the way down to vacant squares.
I remind you that the value of a vacant square is a very fickle thing. It can change without warning – from move to
move, from the tiniest little value to the infinitely great, and then back!
Naturally, the greater these fluctuations, the harder it is to tally up the strongest move. Sometimes this becomes so
difficult that we give up – we acknowledge our helplessness and, almost triumphantly, declare that in the position there
are two (and sometimes more) moves of equal strength!
Most often, such cases are encountered in positions calling for the TC, CP, or TCP algorithms. (See Chapter 5.)
As we conclude this chapter, I could hardly go wrong in saying that we succeeded in filling out all the aforementioned
elements of the defense algorithm with rich chess content. I would add that the Petrosian Algorithm is one of the three
fundamental algorithms. The two others are the Tal Algorithm and the Capablanca Algorithm. And this means that,
having passed through the “Petrosian,” we have finally completed our difficult journey through the spectrum of all
chess attacks and defenses. We began with the game Tal – NN, which lay almost entirely in the extreme right part of
the spectrum. And we finished...
Need I continue?
In Chapter 5, we will focus on the so-called “mixed” algorithms. These are the TC, CP, and TCP algorithms for
searching for the strongest move. The first two of these start out, respectively, from the two well-known (to us) points
on the “t” axis of tcr = 1.25 and tcr = 0.80.
The TC and CP algorithms occupy, respectively, the right and left portions of the spectrum. The borders are flexible,
insubstantial...
Meanwhile, the TCP Algorithm generally has no borders! It’s open, both for attack and for defense. It is irrational, and
it originates at the point where t = 1.00 – that is, at the center of the unified spectrum...
A rhetorical question: Isn’t it time for us to get into Chapter 5 territory? Are you ready?
Chapter 5
Mixed Algorithms, or The TC, CP, and TCP Algorithms For Discovering the
Strongest Move
We have already been introduced to these algorithms, although only superficially to the TCP, more in depth to the TC
and CP. In this chapter, we will deepen our acquaintance with them.
The TC, CP, and TCP algorithms are mixed – or, if you like, derivative – algorithms. They proceed from the
fundamental Tal, Capablanca, and Petrosian algorithms.
That said, the TC Algorithm is not the narrow line of the spectrum close to the point tcr = 1.25. It’s something wider – to
be precise, the part of the spectrum ranging from t = 1.00 to t = 1.50. The borders are unclear; they are defined very,
very inexactly.
The main thing about the definition of the TC Algorithm is not its borders, but that the intensity of the attack on
occupied squares is comparable to the intensity of the attack on vacant squares. In other words, when attacking
according to the requirements of the TC Algorithm, we make no distinction between material targets and empty
squares. Why?
Because the chess value of the pieces and pawns that we attack is comparable to the chess value of the vacant squares
being attacked. As a rule, the value of a square occupied by a piece or a pawn is greater than the value of an empty
square. But then, the number of empty squares is almost always greater than the number of material targets!
The CP Algorithm, on the other hand, is the child of the Capablanca and Petrosian algorithms. This antithesis of the TC
Algorithm occupies the area of the spectrum from t = 1.00 to t = 1.00/1.50. A mirror image!
Of course, the edges of the CP Algorithm are also blurry and very inexact. And the main thing is not the edges
themselves, but the fact that, with the CP Algorithm, the intensity of the defense of the empty squares is comparable to
the intensity of the defense of the occupied squares. In other words, when defending according to the requirements of
the CP Algorithm, we make no distinction between vacant squares and squares that are occupied by our pawns and
pieces. Why?
You know the answer!
The TC and CP algorithms meet at the point t = 1.00, and then go off in different directions.
A simple question: is there any great difference between vacant squares in attack or defense?
No.
A more complex question: is there a major difference between occupied squares? Or: is there in general any sort of
major distinction among all the interacting white and black pieces?
Rather than responding directly to this, but with a sufficiently clear hint: I believe in the unified and indivisible nature of
the game of chess.
By definition, the TCP Algorithm is the “sum” of the fundamental algorithms – that is, “Tal” + “Capablanca” +
“Petrosian.” Or the TC Algorithm + the CP Algorithm (which amounts to the same thing). I add that the TCP
Algorithm is the most complex of all the algorithms. In it, the concepts of attack, defense, and strategic play are knit
together into one. We close this chapter with four games on the theme of the TCP Algorithm.
And now for the first of three games on the TC Algorithm. Ready?
No. 31: Spassky – Foguelman
Mar del Plata 1960
10.Qe2
This move looks more like “contrary to” – White gives up a center pawn. A move according to “Tal,” and not the TC
Algorithm. Or is it?
Honestly, nobody knows the exact boundary between these two brother algorithms...
We won’t quibble. And we won’t nitpick over “perfect” definitions and “exact” borderlines. Better that we give our
position, and ourselves as well, one simple question: is the sacrifice correct?
The answer to that question can only be found in a dozen difficult variations...
An alternative to the text move was the quiet and more natural 10.dxe5, and if 10...Qa5+, then 11.c3 Qxe5+ 12.Nfe2
∆13.Bf4 (according to Spassky). In this case, “...White, thanks to his better development, has the better prospects.”
Spassky’s variation is endorsed by Rybka. But that powerful program does not disapprove of 11.Kf1, either; nor of the
distinctly non-human 11.Qd2, which proposes to meet 11...Qxe5+ with 12.Kd1 … 13.Re1. Bravo, Rybka!
Sample variations:
a) 11.c3 Qxe5+ 12.Nfe2 Nbd7 13.0-0 Bg6 14.Re1 0-0-0 15.Bf4²;
b) 11.Kf1 Qxe5 12.Nd3 Qd4 13.Bb3 Nbd7 14.Qe2+²;
c) 11.Qd2 Qxe5+ 12.Kd1 Be7 13.Re1 Qc7 14.Qe2²/=.
The plethora of variations is an indirect sign of a position with prospects.
Spassky continues: “It’s interesting that on 10.dxe5, my opponent, as he said after the game, intended to reply
10...Qxd1+ 11.Kxd1 Ng4. However, in that case, White, by continuing with 12.Re1, would have an advantage. For
instance, 12...Nxf2+ 13.Ke2 Ne4 14.Nxe4 Bxe4 15.e6. Or 12...Bc5 13.Re2 Nxf2+ 14.Ke1 Ng4 15.Ngh5, and if
15...0-0, then 16.e6.”
Rybka adds: 12...Nxf2+ 13.Ke2 Ng4 14.e6 f6 15.Bd3 Ne5 16.Bxh7 Rxh7 17.h5±.
Rybka also adds: 12...Bc5 13.e6 f6 14.Nfh5 g6 15.Ng7+ Ke7 16.h5±; or 13.f3 Bf2 14.fxg4 Bxg3 15.e6 f6 16.Rf1 Bxh4
17.Nh5±.
We present our preliminary results: the computer indicates an advantage for White after the quiet move 10.dxe5. I
repeat: an advantage!
After the text move (10.Qe2), Rybka claims Spassky only has compensation for his sacrificed pawn (see below). I
repeat: only compensation – which means, approximate equality!
10...Qxd4
Of course!
11.0-0
The first line of the all-powerful Rybka. Spassky found the strongest move.
11...b5
And here is what you should never play! This b-pawn push is clearly wrong, since in the diagram position Black must
defend, not attack. He had to coil himself up at once into a spring, retreating and gathering all (or nearly all) of his
pieces near the king. The king is in danger, and you shouldn’t take that lightly!
The tenth World Champion, criticizing Foguelman for his last move, writes: “Instead of 11...b5 Black should have
played 11...Nbd7, when White intended to continue 12.Rd1 Qc5 13.Be3 Qe7 14.Nfh5 with decent play for his
sacrificed pawn, as it would be far from easy for Black to complete his development.”
Instead of 12...Qc5, Fritz and Rybka prefer 12...Qb6, continuing, after 13.Nfh5 (Spassky’s move!) with, respectively,
13...Qc7 (Fritz) or 13...Nd5 (Rybka). In other words, Fritz is prepared to suffer for a pawn, while Rybka prefers
absolute happiness as far as safety is concerned. Happiness, with material equality!
Among other things, after 13...Nd5 14.Bxd5 cxd5 15.Rxd5 0-0-0, Black is not one bit worse. In fact, he may actually be
a little better.
As an alternative to this last variation, Rybka gives a draw after 14.Nf4 N5f6 15.Nfh5 Nd5 16.Nf4. So there!
What does all of this mean?
Answer (in reply to the question posed in the notes to 10.Qe2): Spassky’s pawn sacrifice was not completely correct.
10.dxe5! was stronger.
After 11...b5 Black’s position slowly, but surely, goes downhill. The b5-pawn is a juicy target for attack. Now
queenside castling is simply unthinkable!
12.Bb3 Bc5
And this move is also not right. 12...Nbd7 was better, as at least Black would not be losing by force.
If 14...Qf8, then 15.Nd3 is very strong (Rybka’s recommendation). Continue the analysis on your own.
15.Bxc5
15.Nfh5 is stronger still, when 15...Nxh5 16.Nxh5 Bxe3 loses quickly. After 17.Nxg7+ Kf8 18.fxe3 Kxg7 19.Qg4+
Bg6 20.Rxf7+ Qxf7 21.Bxf7 Kxf7 22.Rd6, Black is beyond salvation. 15...Rg8 is stronger, but here too, after 16.Bxc5
Qxc5 17.Nxf6+ gxf6 18.Qh5 Qf8 19.Rfe1, the computer predicts a win for White.
15.Ngh5 doesn’t look bad either: if 15...exf4 (Black falls into a trap), then 16.Bxc5 Qxe2 17.Nxg7#!. Pretty, isn’t it?
15...Qxc5 16.Nfh5
16.Ngh5 is good, too. Now 16...Nxh5 17.Qxh5 0-0 loses instantly, as 18.Bxf7+! is decisive. I’d like you to extend this!
Rybka holds the position somehow after 16...Nbd7 17.Nxg7+ Ke7±. There’s apparently no forced win for White – but
everywhere it’s ±, ±, ±...
19.Rd3
The second point of the Tal Algorithm (and it’s the first point of the Capablanca Algorithm) – White intends to transfer
the rook on the f-file to the d-file. However, the straightforward 19.Rfe1 also looks plausible, for instance 19...Kh8
20.Nf6 (… 21.h5+-) 20...Kg7 21.Qg3. Now on 21...Kxf6 White wins with the effective 22.Rd5. Extend this line!
19...a5
Black has no satisfactory defense. What kills him are his poor king and the impossibility of bringing the undeveloped
queenside pieces into the game.
20.Rfd1
20...Ra7
“The threat to win the bishop after 20...a4 was illusory, since White would continue 21.Bxf7+ Kxf7 (or 21...Rxf7
22.Rd8+ Rf8 23.Qe6+, and mate in two moves) 22.Rf3+ Ke7 (22...Kg8 23.Qe6+ Kh8 24.Rd8!) 23.Rxf8 Kxf8 24.Qf3+
with Qf6 to follow” (Spassky).
21.Rd6 Kh8
22.Nf6
“Now Black has no satisfactory defense to 23.Nxh7 followed by 24.Bxf7. For example, 22...Bg8 23.h5 Qb4 24.c4, or
22...Qb4 23.Qxb4 axb4 24.Rd8 Kg7 25.Ne8+, winning” (Spassky).
Or 24...bxa2 25.Nxg6+ fxg6 26.Qxg6 (Rybka’s line). After 26...Qxd6 27.Rxd6 a1Q+ 28.Kh2, there is no defense
against mate.
14.Nd5
This move is clearly in “Tal” style, and it’s probably the strongest.
After 14.Nb1, Black has a comfortable game:
a) 14...e5 15.Be3 Nc5 16.Nd2 0-0;
b) 14...0-0 15.Nd2 a5 … 16...Ba6.
Black does not stand worse in either variation.
14...exd5 15.Bxg7
“Here it would be a serious mistake to play 15.exd5??: 15...Qxd5 16.Bxg7 Qxh1 17.Re1 Ne5 18.Bxe5 dxe5 19.Qxe5. It
looks like White is winning, since he threatens 20.Qxe7# and 20.Bb5+ and Qxh8+, but... Black has not yet used his
right to castle, and he unexpectedly wins – 19...0-0!” (Karpov).
15...Rg8 16.exd5 Qc7 17.Bf6
A “Tal”-style move – an open attack on the enemy bishop and, at the same time, a direct attack on the king.
However, White also had another move available – 17.Bd4 (Rybka). This move is clearly in “Capablanca” style. And it
may not be one bit inferior to the game move!
17...Ne5
“The only move, since 17...Nb6 loses after 18.Re1 Nxd5 19.Bg2, and on 17...Nc5 – then 18.Re1 Ra7 19.Bh3 Bxh3
20.Rxh3, and White ties the black bishop up already with a ʻtriple knot’ (by playing Rh3-e3). The last variation features
an amusing helpmate: 19...Kf8 20.Bxc8 Bxf6 21.Qe8+ Kg7 22.gxf6+ Kh8 23.Qxg8+ Kxg8 24.Re8#” (Karpov).
Rybka extends the then-World Champion’s analysis: in the variation 17...Nc5 18.Re1 Ra7, instead of 19.Bh3 the bishop
capture looks to be just as good: 19.Bxe7, and then 19...Qxe7 20.Qd2 Be6 21.f4. As we can easily see, Black loses in
all variations.
We get the distinct impression that Karpov’s attack (an attack on material targets!) is close to its climax.
With this move – the next move, the sixth consecutive move in “Tal” style – White attacks the enemy pawn with his
own pawn.
19...Bf5
Impressive – Black begins a counterattack (19...Bf5 … 20. ..Rc8 and 21...Qxc2+-+). One could also say (without
exaggeration) that Dorfman confronts the white “Tal” of Karpov with his black “Tal”!
The tension grows... The position we are studying – after 19.f4, 19...Bf5, and the overwhelming number of those that
follow – begins to spiral out of control. We teeter unsteadily on the edge of a precipice – chaos grows and presses on
us...
What’s to be done?
Before revealing the twelfth World Champion’s astonishing reply, I note that after 19...Bf5 Black has significantly
increased the total mobility of his pieces, jumping from 43 possible moves to 49 – count them! And that fact, of course,
is most gratifying to Black. But...
But it is equally true that Dorfman, by putting his bishop on f5, has sharply “elevated” his position. The jump in the
third parameter was +3/11, and for Black this comes with a minus sign, since he (the defender) cannot increase his
∆(move) without doing harm to himself.
The “plus” and the “minus” have joined together! How to weigh them against each other? Was Dorfman correct? Was
the bishop move a brilliant inspiration, or the purest kind of adventurism?
In chaotic positions like this, sometimes it is very, very hard to find the one and only strongest move. In such super-
complex positions I advise you, dear reader, not to trust too much even in powerful chess engines. Even they can find it
very difficult...
Question: So what shall we, ordinary flesh-and-blood chessplayers, do?
Answer: We immediately and fearlessly “turn on” our fantasy and select, from the multitude of appealing moves, that
“one and only” move that we like more than any other. Believe in your intuition and in your passion for chess. There is
ecstasy in battle!
After making this choice, you must cold-bloodedly calculate the variations. Don’t calculate them too deeply – two or
three moves deep, at the most, only so as not to blunder and overlook some tactical blow. There!
What does Rybka have to say?
It, too, plays a bishop move, but a different one, 19...Bb7. Our silicon assistant advises us to play more modestly – in
“Petrosian” style, not in “Tal.” Is Rybka correct?
Answers to this and other questions richly scattered about by me, will not be forthcoming. There will only be hints,
recommendations. And amazing delight at the way both sides played.
Chess is truly inexhaustible...
In chaotic positions, sometimes it is very hard to find the one strongest move. In such super-complex
positions, turn on your fantasy and select, from the multitude of appealing moves, the single move that you
like more than any other. Believe in your intuition and in your passion for chess!
In this position (after 19...Bf5), we have: m < 1, t = 35/49 = ~0.71, a bit of superiority (how much?) in the safety factor,
∆k > 0, and approximate parity in the fifth factor.
Diagnosis: “Capablanca”? Or perhaps even the TC Algorithm?
The ambiguity in our diagnosis is due not just to its third, but alas, also to its first factor. Why?
Because White can immediately remove the enemy pawn on e5. With that, obviously, not only the first, but also the
fourth and fifth parameters of the position are set in motion.
The game continued
20.Bh3
and, as Karpov annotates, “...this move gave rise to various arguments.” Arguments? But not for us! Why?
Because in this position there are signs of the Capablanca Algorithm. And “Capablanca,” as we know, is not opposed to
additional exchanges – if ∆k > 0!
An alternative to the text was the natural 20.fxe5. Good move! However, the World Champion rejected it, as he was
afraid of 20...Rc8 21.Rh2 Qa5. Karpov, as he wrote in his annotations, “did not want to play an ending of the sort that
would arise after 22.Qxa6 Qxa6 23.Bxa6 Rc5.” And further: “I foresee the retort: you say, it’s not necessary to go in
for an ending, for there is the move 22.Qf3 (on 21...Qa5), attacking the bishop on f5 and winning a valuable tempo. But
then the struggle flares up again with renewed force: 22...b3! 23.Qxb3 (forced, since 23.Qxf5 even loses: 23...bxa2
24.Qxc8+ Bd8) 23...Rg6.”
But this is where there is a clear blind spot in Karpov’s pre-computer age analysis: after 24.Qf3 Bd7 25.Rf2 (indicated
by Rybka), the struggle is over – because White wins! I ask you: continue this variation!
If we can believe Rybka, then stronger than the “awesome” 21...Qa5 is the other queen move, 21...Qd7. True, then also,
after the likely 22.Qxa6 Rc5 23.Rf2 Rg6 24.d6, White stands much better, but there is no forced win.
It’s good that, while defending, Rybka “presses” the queen close to the king – 21...Qd7. In connection with this, the rook
move (20...Rc8) may also be considered, not a “Tal” move, but a “Petrosian” move. Black instinctively increases the
local density of his pieces around his king – “Petrosian” is “Petrosian”!
22.fxe5
22...Qc4 23.Rdd3
Karpov offers no comment here, probably because he simply overlooked White’s truly non-human move 26 in the
variation 23.Qf2 Qxa2 24.d6 Qa1+ 25.Kd2 Qxb2...
And now: 26.Ke1. Bravo, metal monster! Bravo, Rybka!
Now the “crushing” 26...Rxc2 is very bad: after 27.d7+ (a surprise check) Black cannot avoid mate within a few moves.
So he must content himself with 26...Rc6, covering the b6 square, but then White plays 27.dxe7 with advantage.
Chess is an infinitely complex game, don’t you agree?
23...Qf4+
In his commentary, Karpov does not hide his admiration for his opponent’s play. The World Champion thinks that the
maneuver ...Qc7-c4-f4, followed by putting the rook on c4 (see Black’s next move) “does honor to the resourcefulness
of I. Dorfman.” And further (we are quoting the twelfth World Champion from his One Hundred Victories): “Besides
this check, it was necessary to examine at least two other
possibilities:
1. 23...Rxg5 24.hxg5 Qxa2 25.d6 [and so on, with – to put it mildly – not all error-free pre-computer variations –
A.Sh.];
2. 23...Qxa2 24.d6, and the black rook on the c-file cannot escape the white queen’s pursuit.”
And now, we extend the two lines given by Karpov, with Rybka’s help:
1. 25...Bxg5+ 26.Rhe3 Rc5 (26...Rc4 27.d7+ Kd8 28.Qf3+-) 27.d7+ Kd8 28.Qf3 Bxe3+ 29.Qxe3+-;
2. a) 24...Rc4 25.Qf3 (25.dxe7 Kxe7 26.e6 also wins, as does 25.e6) 25...b3 26.d7+ Kd8 27.Qa8+ Kc7 28.Qa7+ Kc6
29.Rxb3 Qa1+ 30.Kd2 Qa5+ 31.c3+-;
b) 24...Rc5 25.Qe3 Qa1+ (25...Qc4 26.Qf2 Bd8 27.d7+ Kf8 28.b3 Qb5 29.Rhf3+-) 26.Kd2 Qa5 27.e6 Bxd6 28.exf7+
Kxf7 29.Rf3+ Kg7 30.Qe6+-;
c) 24...Rc6 25.e6+-.
The patient reader, I hope, will take these variations to the end by himself without assistance from a computer. They
were all sweated out by Rybka after long hours of analysis – and, to be honest, I’m not totally sure how solid they are...
Chess is a very complex game, for computers and even more so for flesh-and-blood chessplayers...
24.Kb1 Rc4
Brilliant!
25.d6
Good move.
28.Qf3
28...Qxg5
But here this move – in “Tal” style, not “Petrosian” – cannot stand up to criticism.
In the diagram position, Black (the side we are playing) has m > 1, t = 33/37 = ~0.89, a clear “minus” in the safety
factor, and... “Petrosian.” The diagnosis is simple! And with “Petrosian” we are obligated, roughly speaking, to quietly
retreat and pack ourselves in.
It’s hard to believe, but Rybka confidently holds the position after the seemingly crazy 28...Bd8. The machine fears
neither 29.e6, nor 29.d7+. Possible variations:
a) 29.e6 fxe6 30.d7+ Kxd7 31.Qb7+ Ke8 32.Qc6+ Kf8 33.Qd6+ Kg7 34.Qe5+ Kg6 35.Qxe6+ Kg7 36.Qe5+ Kg6
37.Qe6+, with perpetual check;
b) 29.d7+ Kxd7 30.Rd3+ Kc8 31.Qa8+ Kc7 32.Qa7+ Kc8 33.Qa8+, with a draw.
Can this be right?
29.Re1
“White would retain a solid advantage after 29.Qc6+ Kf8 30.dxe7+ Qxe7 31.Qh6+ Rg7, but I wanted more, so I was
satisfied with a quiet (and at the same time, adventurous) move”
(Karpov).
29...Qg2
“This wild game affected my opponent, too, when he had at his disposal a move which definitely was worth considering
– 29...Qg4!?. ʻFrightened,’ White could have fled to a drawn harbor with 30.Qc6+ Qd7 31.Qxd7+ Kxd7 32.dxe7,
when developing his initiative would have required entering into either of two variations (Karpov)...” And he goes on
to give those two lines –
We do not need to see those variations. They sidetrack us from our objective. Our goal, I remind you, is the TC
Algorithm. And to this end, I will henceforth unceremoniously cast off everything that might interfere with its
attainment. Therefore, later – both in this game and in the one that follows it – we will significantly reduce the
collection of concrete and super-convincing variations. The truth – and it is always tied closely to the strongest move in
the position under study – is hardly suffering here!
I will not mince words: here before you, dear reader, is the position that arose after 46 moves:
47.Kb2
47...f6
48.Rf8
48...Qg7
49.Qc8+
10.h4
This h-pawn advance is but one of a number of good moves. “It is hard to say which of White’s moves is the strongest,
but it is clear that my choice does not pretend to rise to it.” And a little later: “... I began to experiment.” (Morozevich,
64, No. 1, 2008).
Morozevich’s move is in “Tal style.” But there are also “Capablanca”-style moves: 10.Nf3 … 0-0-0, or 10. 0-0-0 … Nf3
– playing in the center instead of the kingside.
10...Nd7
“A strange move, by which Black, of his own volition, corks up his pieces; but whether he will munch on the c5-pawn is
still an open question. I mainly considered 10...Nc6 11.h5, while all my engines with one voice insisted on 10...e5
11.h5 f6 12.Bxg7 Nxg7 13.h6 Rf7, and it’s hard to say who’s better” (Morozevich).
I add: one of those engines doesn’t object to 12.Bd2.
So is the “dissident” program right? After all, with 12.Bd2, White refrains from making trades. When you’re attacking
the king, they are pointless!
So how is 12.Bxg7 good? After the h- and g-pawns disappear, White has significantly increased his superiority in ∆k.
Confirm this!
Which is right – the “dissident” program or the “overwhelming majority”? I don’t know!
11.h5
An alternative is 11.b4, and if 11...a5, then 12.Rd1 (Morozevich) when White packs himself in.
11...Qc7
A good reply, as after the queen trade Black is completely freed from the hazards of an attack on his king. However, as
Morozevich points out, the “anti-Petrosian” move 11...Nxc5 is also quite possible. Play might continue 12.Bxg7 Nxg7
13.h6 Qf6 14.hxg7 Re8 15.f4!? d3!? 16.0-0-0 Na4 17.Rd2, leading to “complicated, irrational positions” (according to
Morozevich).
12.Rh3 f5
13.Qxc7
While this move, the Muscovite grandmaster labels a “proper choice.” Why?
Could it be because, up till 12...f5, the players had already drifted into the Capablanca Safety Zone? Yes or no?
Diagnose the position after 12...f5. Note the drift in the “t” parameter!
13...Nxc7 14.Bg5 e5
“A serious, and perhaps better, alternative was 14...Re8, not allowing the transfer of the bishop to d6. The position is
hard to exhaust with variations, but at first sight the bishop on g5 will be difficult for White to put in a good spot: 15.h6
e5 16.hxg7 Ne6 17.Bh6 Nf6 and, despite White’s two extra pawns, I suspect it will be easier to play the position as
Black – his clear central strategy is against the scattered state of the white pieces all over the board ” (Morozevich).
A very important testimonial! Both because it falls from the lips of a participant in the process, and because it is
supported by high-quality computer analysis. Check it out!
17.Rd1
“The critical moment. Even now, after home analysis, it is difficult enough to say what is really going on. The decision I
made was not based on calculation, but was purely intuitive. I rejected the most obvious line, 17.b4, on account of
17...a5 18.Rd1 axb4 19.axb4 e4! (taking away the f3 square from the knight), when after, for example, 20.Rb3 Ra2, it is
easier to play Black, and I constantly have to keep in mind ...d4-d3 and ...Nd4 – although, understandably, objectively
speaking the position would be quite playable. On 17.e3 or 17.Nf3 Nexc5, also, there is nothing clear” (Morozevich).
17...Nexc5
Nor does 17...a5!? (Morozevich’s marks) look any worse, and if 18.e3, only then 18...Nexc5, etc., a paean to the chess
engine Rybka. What would we do without it? Morozevich writes: “At a depth of 28 half-moves, my Rybka gives this
top line (because I don’t understand half of them, I give the moves without commentary): 19.Be2 dxe3 20.Rxe3 f4
21.Rc3 Ra6 22.Rc2 a4 23.Rcd2, with an evaluation of 0.00.”
Commentary to the commentaries: the position after 23.Rcd2 is compact and very pretty. And it undoubtedly fulfills all
the requirements of the Capablanca Algorithm. Could Rybka have learned to play strategically?
18.f4
18...exf4 19.Nf3 d3
Nor is 19...Na4 inferior. Here is one of the dozens (maybe hundreds) of possible variations: 20.b3 Nc3 21.Rxd4 Ne4
22.Bxf4 Ndc5 23.Nd2 Ne6 24.Nxe4 Nxd4 25.Nd6 with full compensation (maybe more) for the sacrificed exchange.
20.h6
20...g6
I’m deliberately refraining from offering variations. The struggle around moves 21-24 is like a pistol duel in a darkened
room...
The smoke has cleared. Here’s what we have: m = 1, t = 36/26 = ~1.38, and a “micro-plus” in the safety factor. The
diagnosis is simple: “Tal,” just “Tal”!
25.Re3
This is no exchange – more exactly, it is not just an exchange, as on 25...Rxe3 there comes 26.Rd8+ (open attack on
Target No. 1) 26...Kf7 27.Bxe3+-.
25...Ne4 26.Bf3
“White strengthens his position as much as he possibly can before taking on e4” (Morozevich). Does the grandmaster
explain himself to us in the pure language of “Capablanca”? Yes or no? Answer, please!
Our answer is: yes, yes, yes!
I add that: from “Tal” to “Capablanca” there is but a single step...
28...Rac8 29.b3 a5
A final word by Morozevich: “Out of practical considerations, he should have ʻpulled out’ 29...b5, so as not to later
suffer grievously, although objectively this sacrifice is insufficient.”
White successfully combines open attacks on material targets with the steady improvement of his own position.
35...Re6 36.Rd8
Further increasing the ∆(move), and also uncovering a threat to the target on h7: 37.Rh8.
36...g5
Agony.
The TC Algorithm is no doubt an aggressive one. It’s the child of the Tal Algorithm and the right (attacking) branch of
the Capablanca Algorithm. So it would be good for us to strictly define its left edge as the point on the “t” axis with the
coordinate t = 1.00.
With t < 1.00, the “attacking Capablanca” turns into the “defensive Capablanca.”
There is no border on the right! And if there is, then it’s a very, very fuzzy one; and recall that at the very beginning of
this chapter we defined it as the point where t = 1.50. Why there, precisely?
I reply: coming from the considerations of symmetry – the numbers 1.00 and 1.50 are equidistant from the “sacred”
number 1.25, signifying the actual demarcation between “Tal” and “Capablanca.”
In practice, the TC diagnosis is always a consequence of our limited, human abilities. Even world champions sometimes
can’t select the strongest move from several candidates – chess was, is, and will be a very complex game! Thus there
arise the mixed algorithms and unclear diagnoses.
My advice to the practical chessplayer: if your position requires the TC Algorithm, then attack vacant and occupied
squares equally. In other words, when attacking in these situations you have no right to be just “Tal” or just
“Capablanca.”
And what do you do if the position requires the CP Algorithm?
Instant answer: defend vacant and occupied squares equally. Defense is the opposite of attack...
With the TC Algorithm now behind us, the CP Algorithm awaits. Forward!
3) rough equality in the safety factor, since White’s great kingside pressure against his opponent (Qd2 + Be3) is
balanced, not only by the black light-squared bishop’s pressure on the h3-pawn, but also by Black’s “extra” piece in the
f1-f7-h7-h1 rectangle, plus the greater local piece density around his king, relative to that of his opponent.
The preliminary verdict: the position demands the CP Algorithm.
The fourth factor: ∆k < 0.
The fifth factor: ∆(13...Bd7) = 37/16 – 38/16 = -1/16 = ~-0.06.
Our final verdict: the CP Algorithm, where in any case the “Petrosian” slightly predominates over “Capablanca.” The
reason for this barely perceptible drift from the almost “100%” CP Algorithm over to the Petrosian Algorithm is
obvious: ∆k < 0. Our opponent is more compact!
What to do? The answer is given by the World Champion:
14.g4
Why?
Because, “...I didn’t want to play the immediate 14.f4” (Karpov). And further, “It would be illogical to increase the
tension right away – you would have to play g3-g4 eventually anyway... and Ne2-g3.” And now for perhaps the most
important part of Karpov’s verdict: “At the same time [after g3-g4 and Ne2-g3 – A.Sh.], White also completes an
important strategic task: he shores up the e4-pawn.”
While shoring up his central pawn – that is, protecting it – Karpov also increases his ∆(move). But he accomplishes this
with no loss in compactness. How does he do it?
I reply: Karpov does not “elevate” all of his pieces and pawns randomly, but only those that are located in the middle of
his battle formation. And those are: 14.g4, 15.Ng3, and 16.f4 (forgive me, I’m getting a bit ahead of myself – see
below). What matters to us is that the advance guard of the position (the d4- and e4-pawns) is kept in place. Karpov
plays carefully and doesn’t rush to cross the border...
A quick analysis of the diagram position gives us: m = 1, t = 37/35 = ~1.06, approximate parity (or maybe a “micro-
minus”) in the factor of safety, ∆k < 0, ∆(16.f4) = 41/16 – 37/16 = 4/16 = 0.25.
Diagnosis: “Capablanca”? CP Algorithm? Or: a CP Algorithm where “Capablanca” predominates slightly over
“Petrosian”?
We see that, compared to the position after 13...Bd7, White has drifted slightly to the right over the last three moves
(14.g4, 15.Ng3, 16.f4).
Success? Unquestionably!
16...b5
A move of dubious merit. It might have been better to play 16...exf4 17.Bxf4 Qb6 (Rybka’s recommendation), for
instance 18.b3 Qb4 19.Rf1 Rf8 20.Be3 … Rf2, Rdf1².
17.a3
Karpov is in no hurry and institutes a policy of prophylaxis against his opponent’s intended counterplay on that wing,
where the opponent has superior forces.
The computer at first prefers 17.f5 to the game move. But then, after thinking a bit, it agrees with the World Champion.
Hand-to-hand combat ensues after 17.f5 b4 18.Nce2 d5. One possible continuation would be 19.g5 Nxe4 20.Nxe4 dxe4
21.fxg6+ fxg6 22.gxh6 exd4 23.hxg7 dxe3 24.Qxe3, and here, if White is better, then it’s only by the tiniest amount.
I repeat: Karpov is in no hurry! But Timman is:
17...b4
18.axb4 Nxb4
– the white a-pawn disappears and, with it, White becomes more compact, if only by a little. Black, on the other hand,
has diminished compactness with respect to the king and pawns.
Thus Black has played poorly, and he deserves an exemplary punishment! How?
Through an “endless” series of prophylactic undertakings!
19.Nce2
19...exd4
Now the white c-pawn will never get to d4. But... out of the frying pan, and into the fire!
20.Nxd4 a5
Were it Black to move in this position, he would have no problems at all after 21...Ba6, when 22.c3 is not possible
because of 22...Nd3. However...
21.c3
An open attack on the opposing knight, which White combines with prophylaxis!
21...Na6
22.Qc2
“A subtle move which, first of all, does not allow the black knight to jump to c5 (in view of the reply b2-b4!); and
secondly it continues the previous strategic line – shoring up the e4-pawn” (Karpov).
The impatient Rybka suggests 22.Nxc6, followed by 22...Qxc6 23.e5 d5 24.exf6, with an evident advantage.
22...Bd7
“Still intending to carry out the advance ...Na6-c5, for which it is necessary to defend the c6-pawn ahead of time
(23...Nc5 24.b4 axb4 25.cxb4 Ne6)” (Karpov).
23.Nf3 Re7
24.Bf2
Splendid!
The author of this monumental piece of chess art writes: “One of the last prophylactic moves. Before his decisive
aggressive operations White lays out his pieces in the most harmonious manner and... once again shores up the central
outpost at e4! 24.Qd3 would be premature in view of 24...Bc8...”
No rush!
24...Be8
A tactical blunder, dropping a pawn for no compensation. More stubborn resistance was offered by 24...Be8, defending
the target on a6 (Rybka’s recommendation).
After 24...Bc8 White has many tempting continuations. There’s 25.Nh5, 25.e5, and the stealthy 25.f5. One of the lines,
and I am not sure that this one is best, is 25.f5 Nd7 26.g5 Ree8 27.fxg6+ fxg6 28.Nf5±. I ask that you extend this line,
and without using a computer, as far as you can!
“Tal”? Yes, because t(24...Bc8) = 46/39 = ~1.18. With a small “plus” in the safety factor!
The game move (24...Be8) is bad not only because it abandons the piece on a6 to its fate, but also because after it, the
mobility of the black pieces decreases significantly – from 39 to 33 moves.
All this means that White is guaranteed a successful attack!
25.Qd3
25...Qb7
If 25...Nc5, then 26.Bxc5+-. 25...Nb8 is bad, too: White wins with 26.e5+-.
26.Ra1
“For ʻtotal happiness’ White has only to play c3-c4, to gain complete domination over the entire board” (Karpov).
That’s it!
11.Rd1
The most natural move, and most likely the strongest – White not only defends the attacked pawn, but also increases the
density of the pieces around the king. The king is a piece of infinite chess value!
An alternative is 11.Ne4.
11...h6 12.h3
“White plays the opening without extending beyond the third rank, but in passing he frightens Black with the ephemeral
threat of g2-g4 (the studying of which must have cost Tukmakov more than just the odd minute)” (Kasparov, My Great
Predecessors, Part IV).
12.Be2 (Rybka’s recommendation) looks a little better.
12...Qe6 13.Nd2
Preventing 13...e4.
13...Nd7
“The standard 13...Rd7 14.Be2 Rad8 15.Nce4 Nd5... wasn’t bad” (Kasparov). After 16.0-0 Qg6 17.Kh1 the position is
about equal.
14.Be2
Not the slightest trace remains of Black’s former superiority in the safety factor. White’s king, as we can easily see, is
now more than safely bundled. However, this parity has been purchased at the price of some of the “t” parameter.
14...Kh8
A careful move – and a hard one to understand. But he didn’t need to make it. Black must play actively, since after
14.Be2 the diagnosis is either “Tal” or the TC Algorithm. Figure it out on your own!
The strongest move here is 14...Qg6, as indicated by both Kasparov and Tukmakov himself (in Bobby Fischer, vol. 3,
1993, edited by S. Solovyov), as well as Rybka. A sample variation: 15.Nd5 Bd6 16.g4 Be6 17.Bf3 Nc5 18.Nc4 a5,
when Black is slightly better.
15.0-0 Bg6
“Again a non-concrete move. 15...Qg6 16.Kh2 Nb6 came into consideration” (Kasparov).
16.b4
16...a6 17.Rc1
Clearing d1 for the king’s rook – the first point of the strategic algorithm. White already stands no worse.
17...Rac8
Stronger is 17...Nb6, not allowing the knight on a4 and preventing 18.Nb3 because of 18...Nc4.
18.Rfd1 f5
19.Na4
With this move, White starts an attack on the side of the board where he has an “extra” queen. It is obvious that Fischer
has seized the initiative, and he is truly fearsome when attacking...
19...Na7
Kasparov gives this move a question mark, with the comment: “Ugly moves are rarely good. 19...f4?! 20.exf4 (20.Bf3
fxe3 21.fxe3 Nb6 22.Bxc6! is also possible) 20...exf4 21.Re1! Qf7 22.Bf3 Ndb8 23.Qb3! was dangerous for Black, but
19...Nb6 20.Nc5 Bxc5 21.bxc5 Nd5 would have retained a more or less sound position.”
Kasparov’s maxim is important: “Ugly moves are rarely good.” In other words, we rely on esthetics when the position
becomes too complex for us. In such positions, fantasy and intuition begin to take over for logic and accurate
calculation...
20.Nb3 b6
21.d4
21...f4
Open attack on the queen! And more: the pawn has crossed the border – a declaration of war!
What to do?
Answer the opponent’s illegitimate “Tal” with a real “Tal”! Taking into account the complete parity in all the
parameters of the position, Black has no basis at all for mounting a dynamic attack, and should be punished. The
coming, rapid “Tal” by Fischer is a full-strength “Tal.” Tukmakov has raised his sword, and by the sword he shall fall...
22.e4
22...Nb5
23.Bg4
Of course! The king is unapproachable, so therefore White attacks the second big target in the value scale.
23...Qf6 24.dxe5
We know that when playing according to the requirements of the TC Algorithm, when in the attack we assault both the
occupied and the unoccupied squares in our opponent’s territory at the same time. The zone of the TC Algorithm (in its
broadest application) is the whole right half of the unified spectrum of all attacks and defenses.
By the same token, when playing according to the requirements of the CP Algorithm, when in defense we defend both
the occupied and the unoccupied squares in our own territory at the same time. The area of the CP Algorithm (in its
broadest treatment) is the whole left half of the unified spectrum of all attacks and defenses.
We conclude our introduction to the CP Algorithm with a game by Khalifman against Lékó; this algorithm is the mirror
image of the TC Algorithm. These two algorithms have a common point on the “t” axis: t = 1.00. This point splits the
spectrum in half, but at the same time, it brings them together!
When playing according to the requirements of the TC Algorithm, when in the attack we assault both the
occupied and the unoccupied squares in our opponent’s territory at the same time. The zone of the TC
Algorithm is the whole right half of the unified spectrum of all attacks and defenses.
By the same token, when playing according to the requirements of the CP Algorithm, when in defense we
defend both the occupied and the unoccupied squares in our own territory at the same time. The area of the
CP Algorithm is the whole left half of the unified spectrum of all attacks and defenses.
20.Rf1
White frees the rook from the opposing queen’s stare, and more importantly attaches it to the king. Clearly a
“Petrosian”-style move.
Unfortunately, the equivalently “packing” king move 20.Kf1 offers nothing real. After 20...Nxf3 21.Bxf3 Qa6+ 22.Kg1
Qxb7 23.e5 Qb4 24.Bxa8 Rxa8, Black is more than likely not one bit inferior. With this variation, Khalifman began his
extensive commentary on this game, published in the magazine Shakhmatny Peterburg, No. 2 (16), 2000 – six densely
formatted pages in small type. Hundreds of variations, thousands of moves... And, may I add, annotations of the very
highest caliber!
We will not drown in that sea of variations, since I toss aside all extraneous material. I leave in only a few alternatives to
some of the game moves.
My advice to you is: when playing over a game, turn on your chess intuition, but while doing so don’t go beyond the
bounds of the diagnoses of the positions we study. And don’t let yourself get sidetracked by the analysis of long
variations, since your intuition is dominant. Remember: you can’t encompass the infinite!
We are searching for the next move. We are at the same time both fearless and cautious. We work by a scheme proven
by many generations of chessplayers: we choose a move (the move we like more than the rest), we test it (is there an
immediate tactical refutation?), and then we make our move, not fearing the consequences!
20...Bf8
The alternative was 20...Rd7, with the continuation 21.Rxd7 Nxd7 22.f4. Variations are 22...Rb8, 22...Qb4, 22...g5,
22...Rc8, and 22...e5, with evaluations ranging from² to ±.
21.d7 Qa2
22.Bb5
22.Re1 Qa5 23.Rf1 Qa2 24.Bb5 a6 25.Bd4 Bg7 26.Bxe5 ½-½ (Kramnik – Kasparov, Linares 1998).
22...a6
After 22...Bg7, possible are both A) 23.Qa4, with unclear play, and B) 23.f4.
Then Khalifman breaks B) down into subvariations B1), B2), B2a), etc. All the variations favor White!
23.Ba4!
Exclamation point by Khalifman, who continues: “This pawn sacrifice was the only chance if White wanted to keep
playing for the win. 23.f4 axb5 24.fxe5 Qa6 25.Rc7 Qa5 (25...b4∞) 26.Rb7 Qa6=.”
23...Nxf3+
23...Nc4 and 23...Qa3 are also subjected to analysis. In the latter variation, Khalifman examines three lines: A) 24.f4;
B) 24.Bb6; and C) 24.Kg2, and then (this was becoming habitual) split the variations into subvariations... all the way
down to C4d22.
24.Kg2
It’s Black to move, and we’ll examine this position from his perspective (which is OK for us – we can invert the board
any time we like).
And so, we have m > 1, t = 42/45 = ~0.93, and... total confusion in the safety factor! Against the “poor” white king we
must balance the “poor,” nearly hopeless rook on d8 – it is doomed, since we can see no defense against Be3-b6 (here
we have made a second-pass evaluation toward a genuine diagnosis of the position – see Part II, Chapter 4).
Which is more important – the “bad” white king, or the “very bad” black rook? Do you have an exact answer?
The final diagnosis: the TCP Algorithm. That is, the TCP Algorithm for Black and the same for White. The position has
gone out of control!
Lékó played:
24...Ne5
The alternative was to check with the knight: 24...Nh4+. This move was subjected to the most careful analysis,
astonishing in its scope. Judge for yourself: before you is one of an enormous number of variations Khalifman
published. We are talking about a variation numbered B4b22322 (phew!): 25.Kh1 Qc4 26.f3 g5 27.Bb6 Bd6 28.Rf2
Be5 29.Qc2 Qb4 30.Rf1 Qe7 31.Bxd8 Rxd8 32.Rb3 Qf6 33.Qe2 Qf4 34.Rb4 Kg7 35.Rc4±:
Position after 35.Rc4 (analysis)
Playing according to the requirements of the position, you – ensconced in the branches of the TCP Algorithm – attain
freedom of choice. You choose between (at least!) two full-valued variations. You will become a creator, an artist. You
will become the co-creator of an artistic production of the chess art. In playing this game, you will create here and
now...
I prefer, not the text move 24...Ne5, but the check 24...Nh4+ followed by 25...Qe4. Why?
Because I prefer aggressive chess. How about you?
25.Bb6
25...Qc4?!
Khalifman’s marks. Other possible moves are 25...Qa3, 25...Be7, 25...Bg7... A veritable mountain of possibilities.
Chaos!
26.Qd4!
26...Qxd4
If 26...Qe2, then White wins with 27.Qxe5 Qg4+ 28.Qg3 Qxe4+ 29.Qf3 Qxa4 30.Rc1 (… 31.Rc8) 30...Be7 31.Rc8
Raxc8 32.dxc8Q Rxc8 33.Rxe7 Rf8 34.Be3 (Rybka) or 34.Rc7 … 35.Bc5 (Khalifman). Extend this!
27.Bxd4 Nd3
Threatening 28...Nc5.
28.Bc6!
28...Nc5
All of the following lose: 28...e5, 28...Nb4?!, 28...Bg7, 28...Bd6, 28...Bc5. And everywhere variations, variations, and
more variations...
Threatening 33.Rc1+-.
32...Rd8! 33.Rd1!
Khalifman’s marks. Here, in this (not easily) winning position for White, we say goodbye to the author of this treasure –
Khalifman was truly outstanding in this game! The winner continued playing flawlessly right up to move 47.
The remaining moves are given without comment – it’s not our theme anymore. The CP and TCP algorithms are long
past, the board holds a practically prosaic “Capablanca.”
33...f5 34.Bxd8 Bxd8 35.exf5 gxf5 36.Rd6 Kf7 37.Rxa6 Ke7 38.Kg3 Bc7+ 39.Kh4 Kxd7 40.Ra7 Kc6 41.Rxc7+
Kxc7 42.Kh5 e5 43.Kxh6 Kd6 44.Kg5 Ke6 45.h4 f4 46.h5 f3 47.Kg4 1-0
Our ideal chess goal is error-free play in all possible positions. The goal is unattainable. Eternal is the desire to attain it.
Our ultimate objective is to have a universal method for finding the strongest move in all possible positions. It’s a
threefold goal, and its three facets are the fundamental algorithms: Tal, Capablanca, and Petrosian. We have made our
way through them in the preceding chapters.
Our present goal is the mixed algorithms, based on filling the “spaces” between the fundamental algorithms. The TC,
CP, and TCP algorithms bring concepts together rather than separating them. They whittle down the difference between
the empty and occupied squares in attack and defense. All that remains is for us to make our way through the TCP
Algorithm and thereby complete the unification process...
Our immediate goal is a head-scratching game by the great Russian player, Alexander Alekhine.
Attention: we are entering TCP territory!
I am sure that you will be prepared for both attack and defense, as the TCP Algorithm is unpredictable. You must be
prepared to play across the whole board. Hard work!
All the empty and occupied squares of the chessboard make up a unified and indivisible whole – the chess position. And
the TCP Algorithm covers everything that happens on these squares...
1.b6
1...Nxd6 2.cxd6
A remarkable position where Black has many possibilities, none of which save him.
If 2...Qg3+, then 3.Kh1.
Other variations (according to Kotov): 2...Rb8 3.b7+, or 2...Rxc2 3.b7+ Rxb7 4.axb7+ Kb8 5.Bxa7+, with mate coming
soon.
Nor is 2...axb6 much better: White wins after 3.Rxc8+ Ka7 4.dxe7, for example 4...Bxe7 (checking with the queen
changes nothing) 5.Bxb6+ Kxa6 6.Ra8+ Kb7 7.Be3+, etc. And 5.Rxb6 is possible, too – recommended by Rybka.
Guaranteed victory!
2...Bc7 looks stronger, but here too, Black is not to be envied. After 3.b7+ Kb8 4.dxc7+ (or 4.bxc8Q+ Kxc8 5.Kh1,
winning – Rybka) 4...Rexc7 (4...Rcxc7 5.Bxa7++-), there is 5.Rxc7 “...creating a curious position. As before, the White
rook cannot be taken, either by the rook, because of 6.Bxa7+, nor by the king, in view of 6.Rc3+; the attempt to take
away the c3 square from the white rook by 5...Qe5+ 6.Kh1 also leads nowhere; in that case taking the rook with either
the king or the queen leads, after 7.Bh2, to a winning ending” (Kotov).
I would add that 5...Qg3+ (after 5.Rxc7) also loses, as then there follows 6.Rxg3 hxg3+ 7.Kxg3 Rxc7 8.Bh2 and Black
is helpless.
In the game, Black played
2...Rec7
10...Neg4
Commenting on the game, Kramnik gave this move two exclamation points and one question mark! This is what he
wrote: “Sailing into severe complications. This sally appears illogical – yes, maybe it is, in fact. In principle, the normal
move is 10...e6, but I didn’t like the potential, or even the immediate 11.f4 Nxd3+ 12.cxd3; and in the Scheveningen,
they usually don’t hurry to make the exchange on d3, which shores up White’s center.”
From the viewpoint of the unified theory of chess, the text move has every right to exist. Why?
Because, located within the scope of the Capablanca Algorithm, Black is entitled to an additional trade (knight for
bishop on e3), as his position is the more compact.
11.Bc1 g6 12.Nb3
White intends f2-f4 and then h2-h3, shutting out the enemy knight on h6.
It’s starting...
14...Nh5
15.Nd5
“An amusing situation: neither 15.h3? Ng3! nor 15.Qf3? Nxh2! will work. But it is possible that 15.Bd2!? was
stronger: 15...Bxc3 (15...0-0? 16.Nd5 Qd8 17.Ba5, and White wins) 16.bxc3 0-0 17.c4, with a position with good
prospects for White” (Kramnik).
15...Qd8 16.Bd2
White seems to have taken the initiative: he threatens the killing 17.Ba5. But...
16...e6
A worthy riposte! Powerful counterplay cannot help but pop up for Black – he hasn’t made a single clear error, playing
in faithful compliance with the requirements of the position.
17.Ba5
Open attack on the queen. Counterattack? Or “Tal” versus “Tal”?
17...Qh4+
18.g3 Nxg3
19.Nc7+
In the midst of this hand-to-hand combat (the TCP Algorithm!), we must make the right choice...
The move chosen by Kramnik, who was then “merely” a candidate for the world championship:
19...Ke7
A flawless move – from the pre-computer age esthetic! However, it is not one bit better than the retreat to d7. And not
only is it not better, it is worse, which I’ll now try to demonstrate.
With 19...Kd7 (… ...Ra8-b8, ...b7-b6), Black targets the wandering knight. Here’s the variation serving that idea:
20.hxg3 Qxg3+ (20...Qxh1+ 21.Kd2 … 22.Nxa8±) 21.Kd1 Nf2+ 22.Kc1 Rb8 23.Rf1 Nxd3+ 24.cxd3 b6 25.Be1 Qh3
26.Rf2 (… 27.Rh2+-) 26...h5, and Black stands better.
Instead of 21.Kd1, 21.Kd2 (playing for the draw) is likely stronger: 21...Qxf4+ 22.Ke1 Qg3+ 23.Kd2 Qf4+, etc.
However, Black can steer away from the repetition with no particular risk by playing 22...Rb8 – and then it’s White
who has problems, after either 23.Rf1 Qg3+ 24.Kd2 Rf8 25.Kc1 Qe3+ or 25.Bb6 Kc6.
In the position following 25.Kc1, the queen trade after 25...Qe3+ is good for Black. Why?
Because the white king is safe, while the black king is not.
On the other hand, after 25.Bb6 he could also take a chance with 25...Kc6, since both kings, obviously, are insecure...
Then could it be that, instead of the direct 23.Rf1, 23.Rh3 is stronger? What does Rybka have to say?
I am going out on a limb, complementing the engine’s answer with two or three provocative questions: 23.Rh3 Ne5 (the
threat was 24.Rf3+-) 24.Qg2 Bf6 (… 25...Bd8; 24...b6 25.Bd2) 25.Rd1 Bd8 26.Ke2 (why?) 26...h5 (why?)... And
another bit of scare tactics: do you fear the check on c5 – 27.Nc5+? Our silicon friend does not: 27...dxc5 28.Bb5+ Ke7
29.Bb6 f5. Extend this! Rybka evaluates this position in Black’s favor. How about you?
Not ruling out the possibility that the best drawing chances exist for White after the very clever move 23.Qg2. In
support of this, a long and not wholly convincing variation – I can’t hide this. One of many dozens – nay, hundreds – of
worthy lines: 23...h5 24.Rh3 Rf8 25.Rb1 b5 26.Bd2 Qf6 27.Ba5 Bb7 28.Qg1 Rfc8 29.Qa7 Rxc7 30.Qxb8 Qf2+ 31.Kd1
Qg1+, with perpetual check.
A somewhat unpersuasive variation makes for a not wholly convincing verdict on the king’s move to d7: after 19...Kd7,
Black is, to put it mildly, not worse!
Am I right?
But in the game, as we know, Kramnik retreated the king to e7. The game continued:
Topalov thinks that he has winning chances, so he rejects variations like 21.Kd2 Qxf4+ 22.Ke1 Qg3+, with perpetual
check.
21....Nf2+ 22.Kd2
“On 22.Kc1 Nxh1 23.Nxa8 Qxf4+ 24.Bd2 Qe5 25.Bc3 Qg5+ 26.Bd2 Qg1+ 27.Qe1 Qxe1+ 28.Bxe1 h5, Black’s
position, with three pawns for the piece, is preferable” (Kramnik).
Rybka offers necessary improvements: 25.c3², 25.Qg2 Ng3 26.Rb1², and 24...Qf2∞. Trading queens favors Black,
since with approximate parity in the safety factor (am I correct?) he – Black! – is more compact. We have slid over into
the Capablanca Safety Zone!
22...Nxh1
If 22...Qxf4+, then after 23.Qe3 Qxe3+ 24.Kxe3 Nxh1 25.Nxa8 Ng3 26.Rg1 Nh5 (26...Be5 27.Nb6 … Nc4±) 27.Nb6
White stands better, as the knight is badly placed on h5 – compare this with the position after 28...h5 in Kramnik’s
variation. However, Black should have probably gone in for this line, because after the game move (see the following
diagram), White has good chances... to mate the opposing king! How?
The first step on this difficult path is 23.Rxh1, and it continues (this variation is approximate): 23...Qxf4+ (23...Rb8
24.Qf1 … 25.Rh3 Qg4 26.Be2+-) 24.Kd1 Rb8 25.Bd2 Qg3 (25...Qe5 26.Bc3+-) 26.Be1 Qf4 (26...Qg5 27.Bh4+-) and
here, not 27.Bd2 with a draw (Kramnik), but 27.Rf1, as given by Rybka. The engine continues 27...Qg5 28.Qf2 f5
29.Bd2 Qh5+ 30.Be2 Qh3 31.Qb6, with a terrific attack. And Topalov didn’t see it...
And so, as the computer analysis shows, after the check on c7 Kramnik’s 19...Ke7 leads to an inferior position for
Black. Is 19...Ke7 worse than 19...Kd7 – yes or no?
At least during the game, it is not possible to answer this question with absolute certainty – chess being so vast...
So what should be done?
Trust your intuition! In complicated positions, this must predominate over the calculation of variations!
In the game, Topalov preferred the rook move over the knight move. He shouldn’t have...
23.Nxa8
Now it’s Black who’s better, and the computer confirms Kramnik’s assessment.
28.Kc3
The strongest move! After 28.Kd1 Nf2+ 29.Ke1 Bf4 30.Bb6 Qg1+ 31.Qf1 Nxd3+ 32.cxd3 Qxf1+ 33.Kxf1 Bd7
34.Nc7 Rc8, Black is clearly better (Rybka).
28...Qe5+
Admirable thirst for battle! Kramnik seeks the whirlwind – and finds it!
Back home, in the peace of one’s study, one can (must!) indicate “...the quiet 28...Qxe2 29.Bxe2 Ng3 with some
advantage” (Kramnik). But during the game, sometimes emotions take over!
I add: Kramnik is no chess aggressor in the mold of the young Tal or the young Kasparov. Kramnik has an instinct for
self-preservation. Rejecting a favorable ending, he crosses no invisible boundaries. He controls his emotions (if only in
part). Kramnik does not offend the position by overplaying it...
This absolutely crazy position probably cannot be plumbed to its full depth. It’s so complicated that even world
champions could drown in it. Even Rybka [which means “fish” – Tr.] – along with all other silicon creatures...
What to do?
You know the answer. Out of two or three equivalent moves, you must select the one that you like more than the others.
And be fearless!
Mistakes are unavoidable. But if you’re afraid of wolves, then don’t go into the woods.
30...Bg7
31.Nb6
“White loses spectacularly after 31.Bb6? Qxb2! 32.Qxg3 (or 32.Rb1 Qxa2 33.Qxg3 Bd7) 32...Bc3+! 33.Kc4 Bd7
34.Nc7 Bd2!! with irresistible threats: on 35.Nxd2 there follows 35...Bb5+ 36.Nxb5 axb5# (indicated by Fritz 3)”
(Kramnik).
31...d5
“Increasing the tension even more” (Kramnik). Incidentally, Rybka considers this to be the only acceptable move – that
happens sometimes. But then, after 31...d5 the machine gives a lot of options – the wealth of good moves is
overwhelming. There are (and this list is incomplete) 32.Qf2, 32.Qe3, 32.Rd1, 32.c3...
Fritz 3’s move, 32.exd5, is inferior: after 32...Qd6+ 33.Nc5 Bd4 34.Nba4 Bxc5+ 35.Nxc5 b6, Black obtains realistic
winning chances. Whereas the “non-human move 33.Kc4!” (as punctuated by Fritz and supported by Kramnik)
probably loses: 33...Qf4+ 34.Kc5 Rd8... Ah, that Rybka!
Does Black win in all variations? Most likely so. Check it out! But if that’s not the case, then I add, seconding Kramnik:
we are truly witnessing non-human play. The superiority of the electronic chessplayer over the flesh-and-blood player
has become overwhelming!
Here’s one of the many very good, possible variations: 35.Rb1 Be5 36.Na8 Rxd5+ 37.Kb6 Kd7 38.c3 Bc7+ 39.Nxc7
Qxc7+ 40.Ka7 b6+ 41.Ka8 Qb7#. Impressive, huh?
32.Ka4
A mistake. After
Black’s threats can hardly be defended against. The dark-squared bishop suddenly joins the attack on the white king. So
then, perhaps, was it stronger to play, not 33.Nxd7, but 33.Kb4 (a recommendation by Fritz 11)?
35.Bb6
“Opening the way for the king. And 35.Nc5+ Kc6 36.exd5+ Qxd5 37.Be4 (37.Qxg3 Qd4+) 37...Nxe4 38.Qxe4 Bf8!
39.Qxd5+ exd5 and 40...Bxc5+ is no cure-all, either” (Kramnik).
35...Qxb2
Kramnik gives this move a question mark and then writes: “Time pressure – too bad! It would have been much cleaner
to play 35...Nxe4!! 36.Bxe4 Qxb2! (36...Qxe4+ 37.Qxe4 dxe4 leaves Black with the advantage – five (!) passed
pawns, but there’s still a lot of fight ahead), with this example of a fantastic finish: 37.Rb1 (if 37.c4, then 37...bxc4!
38.Rb1 Bf8+ 39.Bc5 Bxc5+ 40.Kxc5 Rc8+ 41.Kb6 – mate follows 41.Kb4 Rb8+ 42.Kc5 Rb5# – 41...Qa3, and the
white king is doomed) 37...Rc8!! 38.Bc5 (38.Rxb2 Rc4+ and 39...Ra4#!) 38...Bc3+! 39.Qxc3 a5+.”
I will not tweak this analysis, which was published long ago. I turned off Rybka, and the latest version of Fritz as well –
they will not help us. They don’t help us at the board.
What to do?
I reply:
1) overcome your gut-level fear of uncertainty;
2) search for the strongest move, relying on the universal search method;
3) believe in good fortune...
Caissa loves the strong in spirit, and she sometimes blesses them – the fearless fighters! – with her gifts. She blesses
through a feel for the beautiful. Chess beauty will aid us in our searches...
Four ply (or half-moves) – the first line by the most powerful chess engine.
38.Bc5
38...Bc3+
The TCP Algorithm is the most complicated of all the algorithms we know in the search for the strongest chess move. It
is a puzzle, even for the strongest chessplayers, and that’s a good thing!
The TCP Algorithm is not a mechanical blend of the algorithms of Tal, Petrosian, and Capablanca. It is something more,
unified and indivisible. With the aid of the TCP Algorithm, we stare into chess infinity.
We see chaos, and attempt to reorder it. Sometimes we succeed, more often we do not. We try again. With hope, with
faith...
Intuition reins in accurate calculation. It reins it in, but not entirely. Imagination scoffs at logic – be it linear or even
two-dimensional. The chess universe has at least three dimensions – and, truly, it’s impossible to encompass the
infinite...
Chess is too vast!
The following instructive example comes from one of the most notable games by Peter Lékó, the gifted Hungarian
player and world championship challenger. In this example, you will witness unhurried strategic play, combined with a
straightforward increase in the ∆(move) of White’s position.
A “Capablanca”-themed game? And if so, then where is our TC Algorithm?
The TCP Algorithm lies in the first move, a move of fantastic power and indescribable beauty. And afterward –
movement by inertia, more or less quiet play, dominating over the primeval chaos...
28.c4
“Tal” doesn’t work here! And the threatening 28.Qxh7, and the no less terrible 28.Rxd4, don’t win either! Here’s why...
On 28.Qxh7 Black has the remarkable 28...Qc3:
Position after 28...Qc3 (analysis)
Hard to believe, but true – it’s a draw! Sample variations:
1) 29.Qxg8+ Kd7 30.Qxf7+ Kd8 (30...Kc6 31.Qc4++-) 31.Qg8+ Kd7 32.Qf7+, with perpetual check;
2) 29.Qxf7+ Kd8 30.Qxg8+ Kd7 and so on, as in variation 1;
3) 29.Bxf7+ Kd8 (29...Kd7 30.Be8++-), and now 30.Qxg8+ Kc7 is bad, but 30.Kc1 (Rybka) saves Black. It’s a draw
after 30...Rxg4 31.Be6: 31...Qa1+ 32.Kd2 Qc3+, etc.
4) 29.Rxe5+ Kd7 (test 29...Kd8) 30.Qxf7+ Kc6 31.Qd5+ Kc7 32.Qa5+, etc. Draw!
The idea of 28...Qc3 belongs to GM Amador Rodríguez, and all of the variations presented are his, with the exception
of the sidelines, which are Rybka’s.
But the analysis of 28.Rxd4!? (Rodríguez’s mark) is questionable.
After 28.Rxd4 Qc3 (the standard reply) 29.Rxe5+ we get the following position:
Position after 29.Rxe5+ (analysis)
Here Black is better! Don’t believe it? I’ll prove it to you!
29...Kd8 (29...Bxe5 30.Bxf7++-; 29...Kd7 30.Re7+ Kc6 31.Rc4++-) 30.Rxd6+ Kc7 31.Re7+ Kxd6, and it is clear that
White’s attack has run out of steam – there’s not even a perpetual check! After 32.Qf4+ Kc6 (32...Kc5 33.Rc7+=)
33.Qe4+ Kc5, the “good” checks are over, and White must settle for an inferior ending after 34.Qe5+.
Rodríguez looks only at 33...Kd6 and assesses the position as “unclear.” Rybka improves on this: 34.Kc1 with a long
variation and a general evaluation of (0.00).
One more variation: 33.Qc1. What to do? I say: 33...Rcd8 or 33....Rgd8, either way with the threat of 34...Rd1. White is
in very bad shape. See for yourself!
So if we retreat, dreaming of a draw, then isn’t it better to retreat the queen to c1 a move earlier – 32.Qc1? The threat is
33.Qd1+ and so forth: checks on g1, h1... Draw?
Alas! After the powerful 32...Rc5 (all hail our silicon friends!) Black’s advantage is unquestionable: 33.Re1 Re5
34.Rd1+ Ke6. White’s king is in a mating net, while Black’s stands in relative safety. Who’s attacking whom?
The overall conclusion: neither 28.Qxh7 nor 28.Rxd4 gives White any advantage.
Rybka advises 28.Rd3 (first line). Evaluation: ± after 18 ply. Lékó’s move, 28.c4, is evaluated lower by the machine: =
(0.22) after 18 ply. Who’s right?
Lékó was right, since after
29...Kd8
30.Qxh7 Rf8
In this position our m = 1, t = 35/22 = ~1.59, a small “+”, ∆k < 0 (6.40 < 6/36), and ∆(30...Rf8) = 0.30.
The diagnosis is unequivocal: “Tal,” and no TCP algorithm! Why?
Because the white king is safe. It is securely shielded by the b3-pawn... and by the enemy pawns on c3 and d4. It is
critically important for these pawns, harmful to their own side, to be securely blockaded. The opponent has no
constructive way to get rid of them. Black is doomed!
Because Black is more compact, and because his king is sufficiently well protected, a frontal assault (the first point of
the Tal Algorithm) has little chance of success. The second point of the algorithm for the attack on material targets
(which is analogous to the first point of the Capablanca Algorithm) applies here. Concretely, Lékó transfers his bishop
to g4.
Rybka prefers leaving the bishop on f3 and the queen on e4. What can I say? All roads lead to Rome!
32...Qc6 33.h4
33...Rb4
34.h5
34...d3+
35.Qxd3 Rd4
Or 35...Rxg4 36.Qxd6+ Qxd6 37.Rxd6+ Kc7 38.Rd5, with an easily winning four-rook ending.
36.Qf5 Rd2+
3) he stands better (by how much?) in the third factor, and that “better” is a combination of two vaguely defined
components: a large “+” (how great?), thanks to a very high level of safety for the white king, and a much smaller (but
by how much?) “-”, signifying the safety situation for the hanging knight at a3.
Our preliminary diagnosis: TCP Algorithm!? Yes or no?!
The third factor of the current position forces us to move to the right along the “t” axis, but where to, exactly? Perhaps
to some point on the line between the Petrosian Algorithm and the Capablanca? Into the Capablanca Safety Zone? Or
even into the Tal Zone?
The fourth factor: ∆k > 0, because 5/25 > 4/25.
The fifth factor: ∆(30...Kxf8) = ~2.70 – 3.33 = ~-0.63.
We are more compact, and also less “elevated.” This means that thanks to the fourth and fifth factors of the position, we
have drifted to the right – maybe just a little. We are closer to “Tal”!
Our final diagnosis: the TCP Algorithm. No question about it!
The left edge of our TCP Zone is some point on the approximate border between the Petrosian and Capablanca
algorithms. The right edge, undoubtedly, is right in the Tal Zone. From “Petrosian” all the way to “Tal.” It’s hard
work...
In this very difficult and fantastically interesting position, the thirteenth World Champion found a very strong move:
31.Kh2
31...Rb3
A worthy response! In Kasparov’s opinion, this is the strongest move, a view confirmed by the soulless computer.
Karpov based his choice on his overwhelming advantage in the time factor. Black is required to attack!
It would not be out of place to note other moves analyzed by Kasparov in those days before computers: 31...Qxa3?,
31...Kg7?!, and 31...Qc1?!. Kasparov awards an exclamation point to Karpov’s 31...Rb3.
32.Bxd3 cxd3
“Objectively this move is no weaker than the others, but now Black is faced with problems which are practically
impossible to solve at the board, especially in time trouble. The obvious 32...Rxd3 would have led by force to an
endgame a pawn down – 33.Qf4 Qxa3 34.Nh6 Qe7 35.Rxg6 Qe5 36.Qxe5 Nxe5 37.Rxa6 Rxd5, but where White’s
chances of success would be very limited: 38.Rxa8+ (38.Nf5 Nf3+!) 38...Ke7 39.Nf5+ Ke6 40.Ne3 Rc5... But such a
metamorphosis obviously could not satisfy Karpov. How could it: to be effectively a piece up, and to go into an
endgame a pawn down?!” (Kasparov)
The former world champion goes on to analyze 32...Rxa3 33.Qf4 as “dangerous” for Black and concludes the analysis
with the following verdict, most unpleasant for Karpov: “Black will have to twist and turn in his search for equality.”
We will (largely) pass over Kasparov’s analysis – you cannot take in the infinite. Better we should, with Kasparov,
make the text move –
33.Qf4
– and then let ourselves be tempted. Beckoned by Garry, we shall make three ply, three steps to the edge of the
precipice: 33...d2 (in the game, Karpov played a different move) 34.Nh6 Nf6:
33...Qxa3
34.Nh6
34...Qe7 35.Rxg6
35...Qe5
“For an instant Black has everything in order – White’s strongest piece is crippled” (Kasparov).
36.Rg8+ Ke7
37.d6+
“By sacrificing itself, this weak little pawn not only saves its own queen, but also wins the opponent’s! As if at the
waving of a magic wand, the scattered white pieces achieve amazing harmony” (Kasparov).
Karpov resigned after
To conclude this chapter dedicated to the TC, CP, and TCP algorithms, I would like to mention their great complexity.
These algorithms are mixed, but they do not divide, rather they unite. They force us to think broadly, and push us on to
generalizations.
We met the TC Algorithm back in Chapter 2, when “Tal” occupied part of the Safety Zone. Our original TC Algorithm
was the narrow little part of the spectrum at the point where t = tcr = 1.25.
And then (in the current chapter) we expanded this part. We literally made the TC Algorithm choke down the entire
“Tal” and the whole right-hand portion of the Capablanca Algorithm. I note also the indisputable (for us) fact that
behind this act of intellectual expansion there stands our unconscious striving for the Unified...
The fruit of these efforts on the TC is a grasp of the overall meaning of all the occupied and unoccupied squares of the
chessboard in the attack. Now, thanks to the expanded version of the TC Algorithm, we make distinctions among these
squares only by their chess value.
We got acquainted with the CP Algorithm in Chapter 3, when we left the Capablanca Safety Zone for the “Petrosian
Zone.” Our original CP Algorithm was the narrow band of the spectrum with its center on the point t = tcr = 0.80.
Striving to generalize, we expanded this narrow segment. We literally forced the CP Algorithm to choke down the entire
“Petrosian” and the whole left half of the Capablanca Algorithm. Once again (and recalling the expansion of the TC
Algorithm), this illustrates our unconscious striving toward the Unified...
The fruit of these efforts on the CP is a grasp of the overall meaning of all occupied and unoccupied squares of the
chessboard in defense. Now, thanks to the expanded version of the CP Algorithm, we make distinctions among these
squares only by their chess value.
The following (and for us the final) step is in the direction of the TCP Algorithm. This algorithm, it’s easy to see, binds
the TC and CP algorithms together in a single whole. Put another way, the TCP Algorithm brings to a common
denominator all the vacant and occupied squares in both attack and defense. This version – let’s call it “expanded” – of
the TCP Algorithm aims to seek the strongest move in all possible chess positions.
A lofty ambition! It’s impossible to control the TCP Algorithm without making an intense effort of will. We must break
down the TCP Algorithm into three categories – “Tal,” “Capablanca,” and “Petrosian.” And more – four elements of
the Tal, three elements of... Can I stop now?
The TCP Algorithm binds the TC and CP algorithms together in a single whole.
It brings to a common denominator all the vacant and occupied squares in both attack and defense.
If, working out of the highest esthetic considerations, we were to refrain completely from carrying out acts of analytical
violence on a chess position, then we would be instantly transformed into sterile contemplatives. On the other hand, if –
in the attempt to get to the bottom of chess – we try to lay everything out one piece at a time, what we get then are
useless masses of seemingly unrelated, disparate moves. Millions and billions of possibilities.
It is not possible to be a “100% contemplative” – that is, a pure intuitive. Geniuses of Capablanca’s sort still need to
calculate variations. Nor does chess know any “100% calculators.” Geniuses of Tal’s sort still have to evaluate the
position based on fundamental considerations. The elusive truth – which always correlates to the strongest move – lies
somewhere in the middle, between the two extremes. And this truth is filled with inspiring beauty and harmony.
Chapter 6
The Algorithm Drift Chart and the Search For the Strongest Move
In this chapter, we summarize. The summary of all summaries: a universal method for discovering the strongest move,
and its integral complement – the Algorithm Drift Chart.
I’ll start by gathering together the most valuable lessons from the preceding chapters, which I will then put into two
categories, conditionally labeled “Parameters” and “Algorithms.” Of course, these will contain not just the parameters of
our study positions and the algorithms that serve those positions. There will be something else in them...
Then there will be also the final stretch, and the finish itself. The finish is the aforementioned Algorithm Drift Chart. This
chart is nothing less than the long-awaited crown of all our efforts, which will tell us about the search for the move in all
possible positions. In all positions, without exception!
Are you ready?
The first category is Parameters – the five parameters and the five factors of the chess position that they represent. They
are:
1) the material factor of a chess position, and the parameter “m”;
2) the factor of chess time, and the parameter “t”;
3) the factor of safety in a chess position, and its parameter;
4) the factor of the compactness of a chess position and the parameter ∆k;
5) the factor of spatial expansion and its parameter ∆(move).
By definition, m = M1/M2.
Here, M1 and M2 are the total chess values of the respective positions of the first and second players. The number 1 we
have already, long ago, assigned to ourselves, and our opponent will always be number 2.
Using as an example the starting position in chess, we total up our chess mass M1 and the chess mass M2 of our
opponent. And so...
In this, the most frequent of all positions, and known to us all, it is White to play, and we have White. The standard
question: what do we have?
M1 = (mq + 2mr + 2mb + 2mn + 8mp)1,
Let’s take each of these three cases in turn. Remember that the “m” parameter stands not only for the drift of the
algorithms, it also stands for the chessplayer’s psychological state. The parameter “m” is two-faceted!
In case (a), when m = 1, we “stand in place,” our psychological stance being neutral. Our level of chess aggressiveness is
medium, equivalent to our level of aggression in the starting position.
In case (b), when m < 1, we drift to the right, from the Petrosian Algorithm toward the Tal Algorithm, and our
psychological attitude comes closer to that of the hungry chess wolf. In other words, we have an enhanced level of
aggression.
Finally, in case (c), when m > 1, we drift to the left, from “Tal” to “Petrosian.” In this case, our psychological stance is
that of a well-fed guard dog, protecting home and host. In other words, we have a lowered level of chess aggression.
Case (a) – the situation where m = 1 – see the Algorithm Drift Chart at the top of this chapter.
Case (b) – that’s where m (→) < 1.
Case (c) – is where m (←) > 1.
The first factor of the chess position is of relatively low intensity; the arrows after the “m” indicate the direction of the
drift. The intensity of the first factor doesn’t hold a candle to the intensity of the third factor...
By definition, t = T1/T2, where T1 and T2 represent the total mobility of all the pieces of the first and second players.
Thus, for example, in the starting chess position, T1 = T2 = 4 + 16 = 20, where 4 is the number of possible moves by the
knights, and 16 is the number of possible pawn moves. Therefore, t = T1/T2 = 20/20 = 1.00.
Of course, total mobility applies not just to the knights and pawns, but to all the chessmen.
Note: The “t” parameter defines the starting point – the point where the drift along the “t” axis begins, and on which the
remaining four parameters of the position apply (or fail to apply) pressure.
The point t = 1.00 cuts the “t” axis in two. The right side goes out to infinity, the left one down to zero. Pay attention: for
the sake of convenience, I have chosen different scales: “t” to the right, and “1/t” to the left.
One more thing: two special points on the “t” axis are: tcr(TC) = 5/4 = 1.25; and tcr(CP) = 4/5 = 0.80. These critical points
define, respectively, the right and left edges of the Capablanca Safety Zone; they are the centers of the areas of the TC
and CP algorithms.
For the sake of simplicity, I almost always drop the signs (TC) and (CP) with the tcr. I do this when there is no possibility
of confusion. So don’t be surprised, dear reader, by my “contradictory” meanings for the critical points: tcr = 1.25 and tcr
= 0.80.
What is important, is that tcr(TC) × tcr(CP) = 1.00.
A very capricious, and very sensitive, parameter! Often it cannot be established with any accuracy. It’s a complex
function of the first, second, fourth, and fifth factors.
The “extra” pieces (or lack of them) in the attacking zone, the mobility of the kings and the corresponding local density of
packing – this is an incomplete list of what influences the process of tallying up the third parameter. With this we have to
depend, first of all, on our acquired sense of danger. The usual arithmetic of the other four parameters is of no help here,
alas. In short, the third parameter is not science, but art... And sometimes, it’s all paradox!
White to move
In this position White has m << 1, t = 3/50 = 0.06 << tcr = 0.80, both kings are in a mating net, ∆k > 0, and ∆(move) << 0.
1.g3# is decisive. Mate in the “ultra-Petrosianish” meaning of the “t” parameter!
Of course, the diagram position is artificial, and it would hardly ever occur in a tournament game, and yet...
Even the tiniest “pluses” or “minuses” in the third parameter can add up to a sizable drift along the “t” axis. With a “plus”
in this parameter, we always drift to the right – that is, from the Petrosian Algorithm toward the Tal Algorithm, while
with a “minus” we drift the other way.
And if we have parity in the safety factor, then we basically stand in place. In other words, in this case the third parameter
is switched off – it has no effect.
When it’s not clear which way the third parameter points, we immediately activate “Tal”! And if “Tal” doesn’t work, then
“Capablanca.” At the end of the row of algorithms there stands the Petrosian Algorithm.
The “plus” case corresponds to the “plus” symbol → (see the Algorithm Drift Chart).
The “minus” case corresponds to the “minus” symbol ←.
Parity in the safety factor corresponds to the symbol for “parity” (see the chart).
The fourth factor is a number. In the previous diagram, it is calculated as ∆k = k1 – k2 = 2/6 – 9/32 = 5/96.
Here:
2 = the number of the white chessmen (king + pawn);
9 = the number of the black chessmen (king + eight pawns);
6 = the area of the rectangle f2-f4-g4-g2;
32 = the area of the rectangle a5-a2-h2-h5.
In these rectangles are placed, respectively, all of the white and black kings and pawns.
Three cases are possible:
a) if ∆k > 0, we drift to the right (∆k > 0 →) – that is, from “Petrosian” to “Tal;”
b) if ∆k = 0, we “stand in place;”
c) if ∆k < 0, we drift to the left (← ∆k < 0), from “Tal” to “Petrosian.”
The fourth factor of a chess position is a low-intensity factor, and the arrows next to the ∆k > 0 and ∆k < 0 indicate the
direction of drift. The intensity of the fourth factor bears no comparison to the intensity of the third factor...
The fifth factor of a chess position is also a uniquely defined number. In the above position, it is ∆(move) = 6/2 – 80/16 =
3.00 – 5.00 = -2.00.
Here:
6 = 4 + 2 (the white king is on the fourth rank, and the white pawn is on the second);
2 = the number of white chessmen;
80 = 3 + (5 × 4) + (5 × 5) + (3 × 6) + (2 × 7) (one black piece – the d6-knight – is on the third rank, counting from
Black’s viewpoint; five chessmen are on the fourth; five on the fifth; three on the sixth; and two chessmen – the a2– and
e2-pawns – are on their seventh rank);
16 = the number of black chessmen.
Naturally, there are three cases:
a) if ∆(move) > 0, then we drift left, from “Tal” towards “Petrosian”: ← ∆(move) > 0;
b) if ∆(move) = then we “stand in place”;
c) if ∆(move) < 0, then we drift right, from “Petrosian” towards “Tal”: ∆(move) < 0 →.
The fifth factor is also a low-intensity one, especially when compared to the third factor, and so we indicate the drift
under ∆(move) > 0 and ∆(move) < 0 with smaller arrows.
We have filled the first of our two “baskets” to the brim with the richest concentrate, prepared from the five factors of the
chess position.
An innocent question: why do we need all of these five factors?
The answer is obvious: these parameters help us to “calculate” the algorithm we’ll use to seek the strongest move.
The second basket is “Algorithms.” The fundamental and non-fundamental algorithms of the search for the strongest
chess move are the “Tal,” the “Capablanca,” the “Petrosian,” and the TC, CP, and TCP algorithms.
We begin with the Tal Algorithm, with its radical sort when “t” aims toward infinity, and then we’ll gradually drift in the
direction of the Petrosian Algorithm.
In the Tal Zone, the algorithm of attack on material targets – the Tal Algorithm – rules.
Its four elements are:
1) open and direct attacks on the opponent’s material targets;
2) the optimal arrangement of our pieces on squares suitable for subsequent open and direct attacks on the material targets
in the opponent’s camp;
3) the sacrifice of chess material (we sacrifice material in order to increase the tempo of the attack); and
4) the win of chess material.
Or, to put it more briefly: when attacking in “Tal” style, the stronger side must –
1) attack;
2) put his pieces on good squares;
3) sacrifice;
4) win material.
In high-mobility situations featuring large “t” values, this element of the algorithm predominates. The second element is
dormant. The third and fourth elements are extremely active and very closely intertwined. They work in tandem, under
the name of “non-equivalent trades.”
As the position drifts to the left along the “t” axis, the influence of the first element decreases and the second element
arises from its slumber to become active. Meanwhile, the tandem of the third and fourth elements of the algorithm goes
through changes: it gradually degrades, turning into a strategic exchange.
Near the point where t = tcr = 1.25, we enter the zone of the TC Algorithm; we’ll discuss this a little later (after
“Petrosian”).
The point t = tcr = 1.25 is behind us now, and we are in the Capablanca Safety Zone – more precisely, in its right half. The
right-handed “Capablanca” is for the strategic attack.
What has changed with our entry into this Safety Zone?
The first element of the Tal Algorithm has dwindled, so welcome to the second element of this fearsome algorithm!
The second element of the Tal Algorithm we will call also the first element of the Capablanca Algorithm.
There are three elements to the Capablanca Algorithm:
1) the optimal placement of our own pieces on squares suitable for a strategic attack on vacant squares of the chessboard;
2) pawn advances;
3) exchanges.
Or, more briefly: when playing in “Capablanca” style, we are obligated to:
1) put our pieces on their best squares,
2) advance the pawns, and
3) seek exchanges.
In other words: in the attack, “Capablanca” attacks first of all not the opponent’s pieces and pawns, but vacant squares.
“Capablanca” thus expands the scale of material chess targets: king, queen, rooks, bishops and knights, pawns – and now
also vacant squares whose chess value is less than that of a pawn (less than 1). This is in the first place.
In the second place, “Capablanca” attacks vacant squares in enemy territory not just with pieces, but also with pawns. It
“elevates” the pawns, moving them closer to the queening square.
And thirdly, in the attack, “Capablanca” refrains from “Tal”-like non-equivalent exchanges, in favor of the strategic
exchange. The reason for this is the lower dynamic energy of the algorithm. The right-hand “Capablanca” is a low-
intensity “Tal.”
Moving further along the scale, we soon find ourselves in the quietest of chess spots – the area around the point t = 1.00.
In this point (and this is important) attack and defense come together. “Capablanca” at this point attacks and defends at
the same time! This kind of play is called strategic play or, the strategic attack/defense of vacant squares.
The final formulation of the Capablanca Algorithm looks like this:
1) the optimal placement of our pieces on squares suitable for a strategic attack or defense with respect to vacant squares;
2) advancing the pawns;
3) exchanges.
Continuing to drift leftward along the “t” axis, near the point t = tcr = 0.80 we fall into the sphere of influence for the CP
Algorithm – we will speak of it later, after the TC Algorithm.
And beyond this point we find ourselves in the Petrosian Zone.
There are three elements to the Petrosian Algorithm:
1) the optimal setup of our pieces and pawns on squares suitable for defending occupied and vacant squares from open
and direct attacks;
2) exchange;
3) the sacrifice of material (we sacrifice material in order to slow down the opponent’s attack on our material and non-
material targets);
Or, briefly: when playing in “Petrosian” style, we must:
1) set out our pieces and pawns on their best squares;
2) seek exchanges;
3) sacrifice material.
When defending, we defend our position according to the value scale of material and non-material targets. The main piece
is the king, followed by the queen, etc. At the bottom of the heap are vacant squares of little chess value.
In defense, the first thing we do is to try to make our position more compact, easier to hold; and, preventing tragic (for us,
at least) developments, we almost gratefully hand over to our opponent part of our expanse of space.
Secondly, we look to make trades. We trade our opponent’s active pieces for our own passive pieces. Depending on our
possibilities, we similarly increase the “t” parameter and, more importantly, decrease the potential for attack. We play for
exchanges, even if our ∆k < 0, because the king’s safety, and that of our other pieces, is foremost.
And thirdly, we are prepared even to give the enemy part of our chess material. We give it up if our opponent’s position is
very strong. Try in this to gain some advantage or other. Give up to your opponent, let’s say, a pawn for a greater degree
of safety in your position. Give up a rook for a minor piece. Give up a minor piece for a couple of pawns. Look for
perpetual check, stalemate, theoretically drawn positions with an overwhelming material advantage. Look for
impregnable fortresses. Keep looking... and pray to the high chess gods to protect your king!
A few words about the mixed algorithms:
The TC Algorithm (narrowly considered) is a “Tal” that knows how to attack only pawns and vacant squares, and that
knows how to sacrifice only pawns. However, it also knows what a strategic exchange is!
Analogously, the CP Algorithm in its narrow treatment is a “Petrosian” that “knows how” to defend only pawns and
vacant squares and that “knows how” to sacrifice only pawns.
The TC and CP algorithms may be broadened, respectively, from t = 1.00 to t = ∞ and from t = 1.00 to t = 0. The centers
of gravity of these algorithms (in their broadest treatment) are close, respectively, to the points tcr(TC) = 1.25 and tcr(CP)
= 0.80.
The most complex of the three non-fundamental algorithms is the TCP Algorithm. Its center of gravity is t = 1.00. This
algorithm combines the TC and CP algorithms (in their broadest treatment). It is brave, sometimes to the point of
insanity!
Now we have also filled the second “basket” to the top. We have filled it with the richest concentrate, prepared from the
three fundamental and the three non-fundamental algorithms of the search for the strongest chess move.
A natural and somewhat scary question: what now?
I reply: we sum up all of our previous work. We are at our goal!
Our ideal goal in chess is faultless play in all possible positions. This goal is unattainable. Eternal is the desire to achieve
it.
Our ultimate chess objective is to formulate a universal method for searching for the strongest move in all possible
positions. And we stand at the threshold of that goal!
The universal method for searching for the strongest chess move lives in the Algorithm Drift Chart at the start of this
chapter.
In Part II of this book, we practice the method. For each position presented, you will:
1) Adopt a medium level of chess aggression.
2) Define “m” and adjust your level of chess aggression accordingly.
3) Define “t” and adjust your level of chess aggression accordingly.
4) Calculate the third parameter. Drift. Preliminary diagnosis.
There are two possible cases:
a) “Tal,”
b) any other algorithm (TC, “Capablanca,” CP, “Petrosian,” TCP).
In the case of (a), consider your preliminary diagnosis your final diagnosis, and go to Step (8).
In the case of (b), proceed to Step (5).
5) Define the ∆k.
6) Define the ∆(move).
7) Correct your preliminary diagnosis and allow for the influence of parameters “m”, ∆k, and ∆(move).
8) “Engage,” one after the other, all of the elements of the calculated algorithm. Look for a move.
9) Make the move.
And finally – this farewell is not mine, but Capablanca’s: “Never refrain from a move out of the fear of losing. If you
think a move good, better to make it, without thinking of the result. Experience is the best teacher. Remember: if you
want to become a good chessplayer, you must lose hundreds of games first.”
With these words, Capablanca ends his Primer of Chess. And with those same words, I conclude the first part of my
handbook...
Part II
Chapter I
Simple Examples
16.Rxf8+
Not only a sacrifice, but also a check – that is, an open attack on the enemy king.
16...Qxf8 17.Qa4+
Another check...
Here, unexpectedly for us, IM Dely resigned. Why?
There is a convincing, four-move variation, demonstrating the hopelessness of Black’s position. For the time being, I’ll
show only the first three moves: 17...b5 18.Qxe4 Rd8 19.Qc6+ Rd7 20.Rd1.
With 18.Qxe4, White not only wins back his pawn, but also strikes a blow in two directions, attacking targets on a8 and
e6. Other, more important material targets are currently out of reach.
Let’s go further – 19.Qc6+. With this move, White renews the assault on the king. The king is the most important target,
and it is good to pursue it.
But why not 19.Qxe6+? Because after 19...Qe7 20.Qxe7+ Kxe7 21.Bg5+ and 22.Bxd8, White has only one extra pawn
in a rook ending, and that’s not enough for him. He wants more...
White’s next move is 20.Rd1. A move of exceptional strength, and not just because White brings a piece into play which
so far has been doing nothing. The main thing is that it threatens mate on the very next move – 21.Qxd7#. To repeat –
the opposing king is the most important target for attack!
After 20.Rd1, Black has only one reasonable move: 20...Qe7:
The high value of the “t” parameter and the more than obvious superiority in the safety factor push White into the Tal
Algorithm’s embrace. “Tal” and “Tal” again!
Under the heading of “strategic attack” we will assume the attack on empty squares in enemy territory.
“Strategic defense,” on the other hand, is the defense of empty squares in our
And something else of great importance: after the third element of the Capablanca Algorithm comes the first! Followed
by the second, the third, once again the first, etc. Exchange upon exchange, cycle after cycle...
The Capablanca Algorithm is unthinkable without the fourth and fifth factors of the chess position.
The fourth factor is that of compactness. The measure of the compactness of a position is how densely the king and
pawns are “packed.” Concretely: 7/16 and 7/24, where 7 is the number of chessmen each side has, 16 is the number of
squares in the minimum rectangle a7-a6-h6-h7, and 24 is the number of squares in the minimum rectangle a2-a4-h4-h2.
Now, 7/16 > 7/24. Black is more concentrated, White is more scattered. Therefore Black, as the more compact side, has
the right to make a trade first. The right, that is, but not the obligation. He does have freedom of choice. To put it
another way, in this position the first and third elements of the Capablanca Algorithm are of equivalent importance.
The fifth and final factor of our position is spatial expansion. The parameter corresponding to it, ∆(move), defines the
degree of “elevation” of the position. The parameter ∆(move) is the arithmetical difference in the numbers expressing
the position of the center of gravity of the black and the white pieces. In our case, this number is ∆(26.Kd2) = -3/7 =
~-0.43 (see the postscript to this game).
We are not interested in the process of calculation itself. More than that: to us, even the final mathematical result is
unimportant. What’s important is the trend, or more exactly: it’s as important as life itself that we define correctly the
direction of the expansion, its vector. In other words, we must answer clearly the question: is this a strategic attack or a
strategic defense? Should we “elevate” the king, or retreat it?
To put it briefly: if it’s a strategic attack, then “elevate” (improve) the king! Because we have t > 1, ∆k > 0, and
∆(26.Kd2) < 0. See the postscript to this game.
Our final conclusion: “Capablanca,” or even the TC Algorithm. There’s something for us to attack!
We firmly believe that the strategic attack is always a consistent improvement in ∆(move). As the stronger side, we
invade the opponent’s territory and occupy it.
Just as firmly, we know that in strategic defense we must cede part of our territory to our opponent if he has the stronger
position. We deliberately sacrifice space in order to save the army from destruction.
And so...
26...Kg5
A strategic attack! Rubinstein invades enemy territory without delay. The black king is “elevated,” the target being the
h2-pawn. Of course, before playing 26...Kg5, Rubinstein calculated everything exactly: White has no possibility of a
counterattack.
27.Ke2
27.Kd3 loses. A rough variation is: 27...Kh4 28.Kd4 Kh3 29.Kc5 Kxh2 30.b5 Kg2 31.b6 (31.Kd6 h5 32.Ke7 b6-+)
31...axb6+ 32.Kxb6 h5 33.Kxb7 h4 and Black wins.
Rubinstein has maximally “elevated” his king, which has taken up the best possible square. Thus we have fully realized
the first point of the strategic (or Capablanca) algorithm. What next?
The answer is evident: proceed to the second point of the algorithm!
29...e5
30.Kh1
30.b5 was better, but this pawn move, of course, doesn’t save White either.
30...b5
“Black fixes the white queenside pawns and leaves himself one tempo (…a7-a6) in reserve” (Rzuvaev and Murakhveri,
Akiba Rubinstein, Fizkul’tura i Sport, Moscow 1980).
Model strategic play! Note the esthetically flawless phalanx of black pawns from the e- to the h-file. Except for the h-
pawn, all have been maximally improved.
34.Kh1 g4
Rubinstein proceeds to the third point of the strategic algorithm. And he is correct – it’s winning in all variations.
However, we are not ruling out the possibility that it might have been simpler to play 34...h4 followed by 35...e4. Why?
Because Point 3 of the strategic algorithm is preceded by Point 2.
35.e4
If 35.fxg4, then the simplest win is 35...fxg4 … 36...e4, 37...h4, and 38...g3.
35...fxe4
In this position, White resigned – and not a moment too soon! Had he made another move or two, something terrible
would have happened. Here’s a variation: 39.f4 exf4 40.e5 g2 41.e6 Kg3 42.e7 f3 43.e8Q f2#. Mate to the white king.
And where? In the pawn ending!
I conclude my commentary to Rubinstein’s strategic gem with the technical parameters of the position after 26.Kd2: m
= 1, t = 14/13 = ~1.08 < tcr = 1.25, “=”, ∆k > 0, ∆(26.Kd2) = 16/7 – 19/7 = -3/7 = ~-0.43.
Speaking of the parameter ∆(26.Kd2), 7 is the number of pieces for each side; while the sums 16 = 2 + 2 + 2 + 2 + 2 + 3
+ 3 and 19 = 2 + 2 + 2 + 3 + 3 + 3 + 4 represent the numbers of the horizontal rows on which the black and white
chessmen, respectively, are placed (for Black, starting from the eighth rank as “1”). (When calculating this parameter,
start with the player whose turn it is to move, and then subtract from it the corresponding number for the opponent.)
Our ∆k > 0 and ∆(26.Kd2) = ~-0.43 impel us to the right. The direction of drift, from “Petrosian” to “Tal,” the starting
point t = ~1.08.
We stop somewhere on the boundary between the right-sided “Capablanca” and the TC Algorithm. This boundary is
fuzzy...
Finally, we need to compare the values of ∆(38...hxg3) and ∆(26.Kd2).
∆(38...hxg3) = 22/5 – 14/5 = 8/5 = 1.60 and ∆(26.Kd2) = 16/7 – 19/7 = -3/7 = ~-0.43.
Simple and – I will not hide it – stunning arithmetic: ∆(38. ..hxg3) – ∆(26.Kd2) = ~1.60 + 0.43 = ~2.03. The growth in
∆(move) was simply colossal. That’s what secured Black a deserved and spectacular win.
Resume: flawless strategic play (strategic attack) is always correlated with the deliberate movement of the center of
gravity of our position towards the queening square.
11.Nd4
If “Capablanca” leaves us free to choose between the first and third elements of the algorithm, then “Petrosian” takes
away that freedom, because our homeland is endangered!
Trading our inactive pieces for Black’s more mobile ones, we gradually eliminate the large deficit in the second factor
of the position. We increase our “t” parameter.
White has gotten rid of two of his minor pieces, including the little-used bishop at g2. What next?
14.Rc1
14....Qe6
17.Qa1
17...Qxa1
“Black does not sense the danger awaiting him in the endgame. After exchanging, he is doomed, in the best-case
scenario, to a rook ending with four pawns (e, f, g, and h) versus five white ones (d, e, f, g, and h). He should have
preferred 17...Qe6” (Karpov).
18.Rxa1 Rfb8
“On 18...Rfc8, White would have achieved the advantage by playing 19.Ra6 Rc2 20.b3 Rxe2 21.Rbxb6, followed by
winning the a-pawn” (Karpov).
Now, what do we have?
t = 40/24 = ~1.67. The rest of it doesn’t matter.
As a postscript to this commentary, we give the technical parameters of the positions after 10...Bd7 and 18...Rfb8.
In the former position, m = 1, t = 34/53 = ~0.64, “=“, ∆k > 0, ∆(10...Bd7) = 28/14 – 31.14 = ~0.21.
In the latter position, m = 1, t = 40/24 = ~1.67, “=“, ∆k > 0, ∆(18...Rfb8) = 24/10 – 19/10 = 5/10 = 0.50.
It is easy to see the colossal progress made in the “t” parameter – from 0.64 to 1.67. We have inadvertently intruded
upon “Tal” territory. Starting from the Petrosian Zone and drifting from left to right, we achieved, with 16...b6, a
“pure” Capablanca Algorithm. We attained it thanks to the hidden inner reserves of our starting position – due to our
superiority in compactness. We simply converted a significant portion of our ∆k > 0 into parity in the “t” parameter.
And we came into the Tal Zone only after Black’s serious inaccuracy on move 17.
Our thinking must be flexible. We are not computer software, but biological entities. We are more accustomed to
thinking in images, to seeking out the strongest move by feel. We should not overestimate the value of the formal
logical component of our thought process. Logic merely fills out our thinking in images. It’s not all-
powerful.
White won the game on move 35.
33.Qg7+
This is how to do it! The check on e6 leads only to a repetition of the position.
1...Kxh4
Or 1...gxh4 2.Qg6#.
2.Qh7+ Qh5
34.Rg6+
34...fxg6
Or 34...hxg6 35.Qg7#.
There is no exact “formula” for calculating beauty. It does not exist – nor can it, since the understanding of beauty is
always wide open – that is, incomplete. It is also significant that the criteria for beauty change over the course of time.
The beautiful is absolutely unthinkable without the paradoxical. Paradox in chess is the paradox of the sacrifice, and its
nature is closely tied to the infinite value of the king...
The beautiful is absolutely unthinkable without the paradoxical. Paradox in chess is the paradox of the
sacrifice, and its nature is closely tied to the infinite value of the king.
I hope that you, dear reader, have already had instilled in you the deepest respect for the safety factor. It’s more
important than all the other factors put together, because the king’s value is infinite!
The third factor pushes us, with great power, out of the Safety Zone into the Tal Zone.
Our final verdict: this position requires the Tal Algorithm.
We begin by engaging all four elements of the “Tal.” Recall that the first of them stands for open and direct attacks on
material targets, whose scale of values starts with the king.
16.Qb5+
This wins.
16...Nd7
If 16...Qxb5, then 17.Nf6#.
17.Rfe1
17...Bb4
True, the fearless Fritz considers 17...Be7 to be the best move here. It gives 18.Ned6+ Kf8 19.Rxe7 (… 20.Rxf7+ Kg8
21.Nh6#) 19...Qxd6... Must I continue?
Black resigned here. There is no defense to the multiple threats of 21.Qe8#, 21.Qg7#, or 21.Qxh8#. A terrible crush!
As an afterword, we present the technical parameters of the starting and ending positions:
After 15...g6: m = 1, t = 44/47 = ~0.94, “+”.
After 20.Qe5: m = 1, t = 47/39 = ~1.21, “+”.
Curiously, the “t” parameter starts below 1, and ends slightly above 1 following 20.Qe5. However, this did not prevent
White from quickly and spectacularly winning the game by a direct attack on the king.
Why? Because of the safety factor!
33.Kf4
Karpov “elevates” his king. Getting a little ahead of myself, I note that Karpov, in his maneuverings, does not wander
beyond the rectangle b2-b5-h5-h2. More than that, at some point he succeeds in cutting the rectangle’s area in half!
33...Kf8 34.Kg5
An excellent square for the king. On the other hand, he wouldn’t look any worse on c5. The variation goes roughly thus:
34.Ke3 Ke7 35.Kd3 (but not 35.Kd4, because of 35...c5+ and then 36.Kxc5 Bxf3 37.gxf3 g5-+) 35...Kd7 36.Kc4 Kc7
37.Kc5 Ba8 38.Be2, f2-f3, g3-g4+-.
34...Ke7 35.Be4
Completing his deployment. The first point of the strategic algorithm (the Capablanca Algorithm) is now accomplished.
35...Ba8 36.f3
Karpov “elevates” his pawn – the second point of the strategic algorithm.
36...Bb7
Nor does 36...c5 save him: after 37.Bxa8 cxb4 38.Bc6 b3 39.Ba4 b2 40.Bc2 a5 41.Bb1 a4 42.g4, Black is defenseless.
37.g4
Points 2 and 3 of the Capablanca Algorithm are fulfilled at the same time.
37...Ba8
White’s position is esthetically flawless. One of the reasons for this is its compactness.
24...Ke6
25.b3
On defense, White strives, as far as possible in this position, to take all his pawns off the second rank. This would allow
him to increase his compactness significantly, and he would obtain chances for counterplay and thus for a draw. But
alas, for this he needs time.
25...g5
Strategically unobjectionable, since Black “elevates” his pawn – the second point of the strategic algorithm. But... but
25...d4 wins immediately! Rybka gives 26.cxd4 exd4 27.Kb4 Kd5 28.h4 Rc2 … 29...d3, etc.
Moving this pawn is stronger than moving the g-pawn. Isn’t it because t(25.b3) = 23/18 = ~1.28? Find out for yourself!
Question: is it “Tal”? Yes! That’s what causes it to be a lightning crush.
26.Kb4 h5 27.h3 e4
28.g3 Ke5
29.Kc5 Rd2
Why so passive? Because Spassky has defended heroically and was now threatening 30.Rd1.
Another question: why does White (if only in one variation) get counterplay, which Black cannot but consider? The
answer: compactness!
In fact, with 27...e4 Black lost his edge in the compactness factor. To be precise, he didn’t lose it, but rather converted
it, trading in his compactness for spatial expansion. He significantly improved his pieces and increased his parameter
∆(move). And that pawn on e4 is his pride and joy, his advance guard. It is the prime candidate for queening.
30.Kb4
Alas, White runs out of resources for a successful defense, he has no good squares left. He is in a pre-Zugzwang
situation.
30...h4
Why? Because all of Black’s pieces are ideally placed. That’s in the first place. And in the second place, for the simple
reason that all of the black pawns, which we would have needed to “elevate” without exchanges, have already been
“elevated.” In accordance with the requirements of the Capablanca Algorithm, Black must now enter into the third
point of that algorithm.
31.gxh4
31...gxh4
We already know well that following the exchange – that is, after the third point of the strategic algorithm – the stronger
side must endeavor to repeat the first point of that algorithm.
What does that mean to us, concretely speaking? That the black king is ready to invade on f3 to decisive effect!
32.a6
Desperation.
32...bxa6
Now Black has only to be accurate in calculating fairly uncomplicated variations. The struggle ends with a short
dynamic phase.
35.Rh6 e3
3) The third factor: the white king is safer. In the first place, it’s better packed; in the second place, it can move more
freely, whereas the black king is semi-paralyzed – it can only reach two of the four vacant squares next to it.
Our preliminary and also final diagnosis is the Tal Algorithm. Simply “Tal,” not the TC Algorithm, since the “plus” in
the safety factor is weightier than the “plus” in the second factor. The powerful factor of safety strongly pushes us
sideways into the Tal Zone.
42.g4
Opening the king’s path through g3 to h4. The king is “elevated” – it rushes to help the queen, since one cannot
checkmate with just the queen.
42...Qd2
Black’s counterplay with 42...e4 … 43...e3 doesn’t work: White gives check on g5 and e8 and then removes the e3-
pawn.
The second point of the Tal Algorithm, the first point being check – that is, an open attack on the main target.
45...g6
Here’s what IM Konstantinopolsky wrote about this position (in 1955): “If the black king were to retreat to h8 or to g8,
the dénouement would also be very rapid: White places his king on h5; then advances his pawn g4-g5-g6, creating
mating threats; and if Black checks from e2, he interposes his pawn on f3 and there are no more checks.”
Not without regret, I must point out that chess engines still hold this position (with the king on h5), by moving the queen
to e8:
46.Qf7+
Another check.
46...Kh6 47.Qf6
47...Kh7 48.Kg5
Not just “elevating” the king, but also attacking the g6-pawn.
I hasten to point out the obvious fact that our “Tal” goes hand in hand with “Capablanca.” In attacking the king, White
is continuously “elevating” his position. On the other hand, the friendship between the Tal and Capablanca algorithms
is nothing new for us.
White wins after 49...Qxf4+ 50.Qxf4 exf4 51.Kxf4 Kh6 52.Ke5, etc. I ask you to extend this, and don’t forget to switch
over from the Tal Algorithm to the “pure” Capablanca Algorithm. Why?
Because, after the queen trade, the position is simplified greatly. The dynamic elements practically disappear, and safety
considerations vanish.
One more question: why does 52.Ke5 win, whereas 52.h4 leads only to a draw?
I ask you to answer this independently:
a) in the language of the Tal Algorithm – that is, in the language of concrete variations; and
b) proceeding from the “abstract” – that is, from general theoretical considerations.
Theory and practice are one!
Mate is unavoidable: 51...f3+ 52.g5, etc. Checkmate is always a triumph for the dynamic chess attack. This is the end
product of the Tal Algorithm...
A simple arithmetic of the game just completed: ∆(51.Kh6) – ∆(41...Qe2) = 1.00 – (-0.35) = 1.35. Work this out on
your own!
And here the growth of ∆(move) brought White his victory – here, via mate to the enemy king.
30...Qxg3
From the viewpoint of our science of chess, this is both an open attack on the king and a piece sacrifice. It’s easy to see
that checkmate is unavoidable. After 31.fxg3 Black mates with 31...f2+ 32.Rxc6 f1Q#. Original chess beauty!
A postscript – the technical parameters of the position after 30.Rd4: m > 1, t = 41/45 = ~0.91, “+”.
The first parameter of this position warns us to be careful. It draws us toward “Petrosian.” On the other hand, the “t”
parameter requires “Capablanca.” And only the third factor calls us to the axe.
Why is “Tal” right?
Because the value of the chess piece that we call the king is infinite. Which means that the higher mathematics of chess
play deals with vast numbers. It’s a peculiar mathematics, and its logic doesn’t jibe with our normal logic.
No. 52: Morozevich – Lutz
Biel 2003
3) The third factor of the position moves us further to the right – obviously, the white king is safer. His castled fortress
is in ruins; but fortunately for us he is surrounded by a sufficient number of pieces devoted to its defense, and so he
stands in relative safety. Still, the black king’s position is alarming, as four of the six empty squares around it are
controlled by his opponent.
One more thing: White has an extra queen within the rectangle f1-f7-h7-h1. In other words, on the kingside we have an
extra piece of artillery. So we have more chances to succeed in our attack!
Our final conclusion: the diagram position requires the Tal Algorithm. And we know well what to do. The four
attacking elements, the attacking value scale...
36.Ne5
For now, the black king is out of reach (checking on e5 would be ridiculous), so we must attack the queen!
Among other things, the other attack on the queen, with 36.Rc7, also wins. Investigate this!
36...dxe5
There’s no satisfactory defense. Fritz gives, as the strongest continuation, 36...Rxg5 37.Nxd7 Rxg3 38.Rf7+, etc. Now
it’s mate.
37.Rf7+ Kh8
37...Kg8 holds out for longer.
Curiously, the significant size of the “t” parameter peacefully coexists with the parity in tempi (5 + 1 = 6). The “t”
parameter weighs in favor of the Tal Algorithm, while the arithmetic of tempi militates for the Capablanca Algorithm.
Which one to believe?
We have long since been through this: we must believe in the “t” parameter, because it is closer to the truth.
My advice is not to ignore the arithmetic of chess tempi during the game. Just use them carefully, with adjustments.
Study the mobility of “good” versus “bad” pieces. Thus, for example, in our starting position the white king, bishop,
and knight are “good” pieces, while the black king, bishop and knight are “bad.” Why?
This is the reason why we say “yes” to the Tal Algorithm and reject the Capablanca Algorithm. In that case, we will be
very close to the victorious knight leap to e5.
White to move
What does White have?
1) The material factor: parity.
2) The factor of chess time: we are in the Tal Zone, since the value of the “t” parameter is more critical (t = 43/29 =
~1.48).
3) The safety factor: clear signs of chess chaos! This position – and of this we have no doubt whatever – is irrational, as
attack and defense are curiously intertwined. For instance, White has the enemy queen under attack, while Black is
hitting the rook on d1. More than that – the white king is threatened with immediate mate! In such a position, a lot (if
not everything) depends on whose move it is...
We have already been over this (see Game 44). Read, repeat. While I, by analogy with the game Polgár – Hansen,
render the final verdict: the diagram position requires the TCP Algorithm.
Where does the TCP Algorithm begin? With open, direct attacks on material targets of the enemy in accordance with
our familiar value scale...
1.Qa8
Winning at once, as after 1...Rxa8 2.fxe7, there is no satisfactory defense to 3.Rd8+. We are privy to a brilliant
combination!
It’s not hard to see that for lack of a better target, White hits the enemy rook. In fact, all attacks on the king or queen
would have been either senseless or losing. Therefore, there remained only one thing – to transfer all possible open and
direct attacks to the rook...
Black resigned. 1-0
A short little afterword – the technical parameters of the starting position: m = 1, t = 43/29 = ~1.48, “?”.
And the “?” sign means an irrational position.
38.Ke4
Of course, it has to be this way! The black king is unapproachable for the time being, and we may, with a clear
conscience, say the same thing about the queen on c7. That leaves the bishop: open attack upon it!
But that’s just the beginning...
Geller’s choice is the first move in a victorious deployment of the ivory army: Kh7/Qe6/Be3. Of course, along the way
we will have to drive the black bishop off the a1-h8 diagonal which is so important to him. But we will accomplish this
easily, and then our opponent’s kingside citadel will collapse!
38...Bh2
39.Kf5 Bg3
After 39...Qf7+ 40.Qxf7+ Kxf7, White wins with 41.Bc1 (… Bb2, Ke4-d5-c6-b7) 41...Bd6 42.Bb2 Bf8 43.Be5. Then
White places his pawn on a4 and tears his opponent’s defenseless queenside to shreds.
40.Bd2
40.Kg6 also wins, since after 40...Qd6 41.Qxd6+ Bxd6 42.Bc1 (… 43.Bb2+-) 42...Be5 43.Kf5 and 44.Ke6, Black
cannot avoid dropping material.
“The game was adjourned here, but was not played out, since the variation 42.Qxf7+ Kxf7 43.Ke4 and 44.Kd5 is
obvious” (Geller).
Black resigned. 1-0
3) rough parity in the safety factor, since White’s king is very safe – and White counters the pressure from the two black
bishops on his castled position with his “extra” minor piece in the rectangle f1-f8-h8-h1.
Our preliminary diagnosis: the Capablanca Algorithm.
What next? Call on the fourth and fifth parameters. In brief: for us, they are at parity and approximate parity. Verify this.
Obviously, the diagnosis of the position doesn’t change – for us, “Capablanca” rules.
There followed
14.Nfe5
14...Bxg2 15.Kxg2
Indeed, if prior to the exchange White had his king and pawns on three ranks, now they sit only on two. That’s a
significant achievement.
15...Nxe5 16.Bxe5
Fritz recommends 16.Qxd8. We have no objection, as White’s compactness is much greater than Black’s.
16...Qe7 17.Qb3
17...Rfc8 18.Rc2
18...Qd7 19.Rfc1
19...Bd4
20.Qf3
20...Bxe5 21.Nxe5
Question: Are you sure that after 21.Nxe5 Khalifman will keep playing according to the Capablanca Algorithm? I’m
not at all sure!
21...Qe8 22.Rc7
22...h6
Fritz likes 22...a6 and 22...Rab8 better, but here too, it adjudicates a win for White.
An open attack on the king, and a sacrifice! When playing in “Tal” style, we know that the chief target of the attack is
the king.
26...Kh7
27.Qe4+ Kh8
After 29...Qxb7 Black comes out a queen down: 30.Qh7+ Kf8 31.Qh8+ Ke7 32.Qxg7+ and then 33.Qxb7.
Let us recall the parameters of the position after 13...Bd5: m = 1, t = 42.38 = ~1.11, “≈”, ∆k = 0...
And now for the position after 25...Qc6: m > 1, t = 40/36 = ~1.11 and a huge “plus” in the safety factor.
“Capablanca” is obvious after 13...Bd5, and “Tal” is obvious after 25...Qc6. The third parameter categorically requires
“Tal”!
The reason for the algorithm drift is Black’s mistake on move 19.
22.Nhxf6+
Not just a trade, but also a check – an open attack on the most important chess target.
22...Nxf6 23.d5
23...Nxe4
If 23...Nxd5, then 24.Bxd5 Bxd5 25.Rxd5, etc. Please extend the variation. Start with a check!
24.dxe6
24...f5
Resigning is simpler.
16.Bxh6
A natural, practically provoked, piece sacrifice. This move didn’t surprise me. I was surprised by a different move
suggested by Fritz: 16.Qe2, threatening check on h7. All of a sudden, there comes 16...Re8, with the following simple
variation: 17.Bh7+ Kxh7 18.Rxd8 Raxd8, etc. Fritz considers this variation almost the strongest for Black!
Question for readers and the authors of the program: didn’t Fritz go out of its mind?
And a request: when you give your reply, refrain from using any other chess engine!
16...gxh6
It was probably better to decline the sacrifice. Down a pawn (say, after 16...Qa5), he could still have offered some
resistance. But now, after –
17.Qxh6
– the computer finds no salvation.
17...Re8
18.Bc4 Bd7
If 18...Qc8, then 19.Rd3 … Rg3++-. And after 18...Qc7, Fritz likes 19.Rd4 best.
19.Rd4
19...Bf8
20.Qg6+
White now has a knight and three pawns for a rook – a material advantage. And the attack on the king still rages.
1-0
10...Bb5
11.Nxb5
11...axb5
12.Bxb5+
Why are these trades so good for Black? Because the black rook’s mobility has increased so much, expanding by a
factor of eight! Even better, Black now threatens mate on the move. In other words, White’s strategic blunder has given
Black good attacking chances.
In the diagram position, we can no longer speak of parity in the first three factors. Material equality has been
maintained, but a small “plus” has popped up for us in the second and third factors. Therefore the requirements of the
position after 13...Rxa2 are met, no longer by the Capablanca Algorithm, but by the TC Algorithm.
Futile; this wastes valuable time. Better was 15.c3, or even 15.f5 exf5 16.Qf4 Na6 17.Qxf5 Nc7 (Fritz).
Now the density of packing of the two kings has evened out: 8/35 = 8/35. Black’s position has turned more brittle,
which cannot be good for him, of course. But in return he has added to his second and fifth factors of the position.
Figure out the parameters “t” and ∆(move) before 15...Nc4 and after 16.Bxc4.
The most important thing is that the exchange has brought the pawn from d5 to c4. Drawing nearer to the white king, it
will participate in the attack on it. Thus Black has also added significantly to the third factor of the position.
To sum up, we may say that Black has very favorably converted his superiority in the compactness factor.
17.Ng5 b5
18.Qe3 Na6
With tempo! Now, of course, White cannot play 21.Nd6+ because of the prosaic 21...Qxd6.
21.Qf3 0-0
Black not only tucks his king away in a safer place, but also brings the formerly inactive rook into the attack.
22.f5 b4
Black’s attack is the stronger one. There is no defense – at least, none that Fritz can see.
23.Rhf1
If 23.f6, then 23...Qa7, etc. Extend this variation – start without a computer!
The value scale for attacking material targets – king, queen, rook...
White’s position is a sorry sight. Within the rectangle a1-a8-c8-c1, Black has an “extra” queen and two rooks – and
that’s not counting that awesome c3-pawn!
I conclude our notes on this game with the technical parameters of the positions following 10.0-0-0, 13...Rxa2, and
28...c3.
After 10.0-0-0: m = 1, t = 32/32 = 1.00, “=”, ∆k > 0.
After 13...Rxa2: m = 1, t = 39/36 = ~1.08, “+”, ∆k > 0.
After 28...c3: m = 1, t = 37/31 = ~1.19, and a large “+”...
The drift is obvious – from a “100%” Capablanca Algorithm to the TC Algorithm, and even into the Tal Algorithm. I
note also that after 22...b4, Black’s attack rolled like an avalanche.
A miracle! Now it’s stalemate after either 2...exd1Q or 2...exd1R. Black gets nothing out of 2...exd1B 3.Be4 Bf3 4.Bc2
(a variation pointed out by Palevich). That’s why Lebelt promoted his pawn to a knight:
2...exd1N
According to Fritz, 4...Nd1 (… ...Ne3 and ...c4-c3) is stronger, with problematic winning chances. But why not grab that
little pawn on a2 along the way?
5.Bd5
Why?
5...c3
6.Bxa2
6...c2
7.Bb1
Draw!
An exact “formula” for the beautiful does not exist. It doesn’t, and it can’t, because the world around us is boundless in
its manifestations.
Nor will an exact, all-encompassing model for playing chess ever come to be, because chess is practically inexhaustible.
And that means that, for us chessplayers, our game will be an endless source of delight.
No. 60: Granberg – Gubnicki
corr. 1986
White to move
What does White have?
1) a piece less – that is, Black has an overwhelming material advantage;
2) the Tal Zone, since t > tcr = 1.25 (work this out!);
3) an overwhelming superiority in the third factor, as the white king’s mobility is at a maximum, while that of the black
king is zero (it’s in a mating net).
Conclusion: the position requires the Tal Algorithm.
We start with checks. I see just two: 1.Rxh3+ and 1.Nf7+. I calculate the variations... These checks, alas, do not work.
See for yourself!
Next I move from open attacks on the king to direct attacks on it, and... O miracle!
1.Nd7
1...Bxd7
2.Rf3
There is no defense!
Gubnicki was mated after
3) perhaps a slight superiority in the safety factor (the X-rays radiating from the black queen are rather stronger than the
clutch of senselessly crowded “extra” white pieces in the rectangle d1-d8-f8-f1 shared by both kings).
Our final diagnosis: the Tal Algorithm.
12...e5
This is no innocent play for massive trades. This is an open attack on the targets at d4 and f4. Black attacks immediate
objectives and at the same time improves the mobility of his pieces (the light-squared bishop and, a little later, the
queen).
13.fxe5
No objections here.
13...fxe5 14.dxe5
14...Bxe5 15.Nxe5
But now...
He had to run for the draw at full speed with 15.Qb3, offering a queen trade. After 15...Bf5 16.Qxb6 axb6 17.Ra1 White
is worse, but certainly not lost!
On the other hand, Black’s reply was very, very difficult to find... Chepukaitis probably saw only 15...Nxe5 16.Qh5+
Nf7 17.Qe2+, with complicated and roughly equal play.
15...0-0
Outstanding! Remember that we are within the range of the Tal Algorithm, and “Tal” directs us to attack. The main
target is the king. Black threatens mate on f2!
There is no satisfactory defense. At least, neither Fritz nor Rybka see one. A few variations to support our conclusion:
a) 16.Ndf3 Nxe5 17.Qxd5+ Be6 18.Qxe5 Rxf3;
b) 16.Nef3 Bg4 17.Be2 Rae8 18.Kf1 Bh3+;
c) 16.Nd3 Ne5 17.Qb3 Qe3+ 18.Kd1 Nc4 19.Nxc4 dxc4 20.Qxc4+ Be6 21.Qd4 Qf3+;
d) 16.Qb3 Qe3+ 17.Kd1 Qxe5.
Please extend these variations. Let that be your homework.
Chepukaitis preferred a different move:
16.Qh5
16...Nxe5
17.Qxe5 Bg4
Spite check.
13.Bh6
Why? Because, first of all, “Petrosian” allows us to play for an exchange. And second, because White is forced to make
the trade, given that moving the bishop to c5 or g5 is bad. So we are limited to choosing only between the text move
and the unpleasant exchanges on d5.
Van Wely is justified! He exchanges the dark-squared bishop for lack of anything better.
13...Bxh6
Fritz prefers the clever 13...Bd4 to this trade. This is easy to understand, because after...
14.Qxh6
14...Qa5
A mistake! This pseudo-active move is strategically deficient. Now White obtains, by force, a position with a backward
pawn on h7, which proves to be a most tempting object for attack.
He needed to take the knight with 14...Nxc3 and then play 15.Qg7 Rf8 16.Qxc3 g5, or 16.bxc3 Qa5 (Rybka), when
Black is out of the woods. I would add that, were it not for the fact that Rybka guarantees 15...Rf8, Black would have
had to parry the opponent’s threat with the paradoxical 15...Kd7, or even with 14...Kd7.Why? Compactness!
And now for a more concrete reply: after 14...Kd7 15.hxg6, Black replies simply 15...hxg6, and the rook is well
protected on h8. The f-pawn recapture is much worse, and you know why!
15.hxg6
Of course! White converts his superiority in the factor of space expansion into an edge in compactness.
15...fxg6
16.Bxd5
White has the more compact position, so he’s entitled to trade pieces first.
16...cxd5
Or 16...Bxd5 17.Qxg6++-.
17.Qg7
Bad would be 17.Qxg6+ hxg6 18.Rxh8+ and 19.Rxa8, on account of 19...d4 when Black comes out a piece ahead.
17...Kd7 18.Qd4
You could play this way. Or you could play it like Fritz, i.e. 18.e4 with attack, for instance 18...Rac8 19.Qe5 Qb4
20.exd5 Bg4 21.0-0, etc.
Done!
and White obtained excellent winning chances. Karpov scored the point on move 57.
The parameters of the position after 21...bxc5: m < 1, t = 27/41 = ~0.66. “-”, ∆k > 0, ∆(21...bxc5) = 16/8 – 21/9 = ~0.33.
The parameters of the position after 28.Ra3: m = 1, t = 23/16 = ~1.44, “=”, ∆k >> 0, because 5/8 >> 5/24, ∆(28.Ra3) =
14/6 – 11/6 = 0.50.
Let us compare, let us be amazed, and let us be enraptured!
Over the space of seven moves, some truly amazing things occurred: White progressed significantly in all five
parameters! It’s easy to see that Karpov not only won back his previously sacrificed pawn, but also moved comfortably
out of the Petrosian Zone – which didn’t appeal to him very much – into the victorious Tal Zone. In addition, White
totally liquidated the ghost of an attack on his king that was so frightening, and – most importantly – he outpaced his
opponent fantastically in the fourth and fifth factors of the position.
I advise you to study this strategic masterpiece most carefully. Use your computer with the most powerful engine at
your disposal.
Our slogan is – compactness!
46.a3
Both a sacrifice and a strategic exchange at the same time. That’s harmony!
Once the “ugly” a2-pawn disappears, Black necessarily falls apart in short order. Why?
Answer: the compactness of the white pieces (the a- through g-pawns went to being the c- through g-pawns)...
48...Rgh8
For some reason, Fritz suggests 48...Nf6, capitulating rapidly after 49.gxf6 gxf6 50.Ra5 Rhxg4 51.Qxg4 Rxg4 52.Kxg4
and 53.Nxc5+-.
49.Rb1 Rb8
50.Qe1 Rxg4
It’s all over, because after 52...Rb6 there is an easy, elegant win with 53.Nxc5 Bxc5 54.Rxb6 Bxb6 55.Bxb6+ Nxb6
56.c5:
43.Rh2
Commenting on this game, Karpov (and he is very stingy with compliments) gives this move two exclamation points.
And he is absolutely right!
Now, as we can easily see, White doesn’t fear the check on g3, because the rook securely defends its colleague on h5.
Victory!
Yusupov resigned after
43...Qd7+ 44.f5
We selected the Petrosian Algorithm, not on a whim, but out of necessity. Saving himself from a mortally dangerous
check, with the rook move White instantly increased his local packing density. Having moved to h2, the rook literally
clung to its king, and saved it!
Before 43.Rh2, the density of packing of the pieces around the king was 2/3 (two pieces in the minimal rectangle h3-
h5). Now it is 3/4 (three pieces in the rectangle h2-h5). Though small, the magnitude of the increase was decisive. White
became more compact!
I conclude my commentary on this final game fragment with two statements on the theme of the strategic and dynamic
branches of the Petrosian Algorithm (see also the materials in Part I, Chapter 4). For now, the bare statements.
1) The strategic branch of the Petrosian Algorithm practically coincides with the CP Algorithm of the search for the
strongest chess move. Particularly if you are entitled to trade first, then you have to exploit it immediately.
2) When defending against a dynamic attack on your king, your only real chance for salvation lies in increasing your
local piece packing density, placed close to your king. You must do this at all costs!
Karpov – Yusupov is a game on the theme of defending against a dynamic attack by your opponent. A sharp attack, a
sharp defense. In the unified spectrum of all attacks and defenses, the place of this game (or, more exactly, of the
position after 43...Qf7) is on the far left half of the spectrum (see the Algorithm Drift Chart).
And we started on the rightmost part of the spectrum – see Game 41. That Fischer game was an illustration of the theme
of the sharpest dynamic attack.
Thus we have surveyed the entire spectrum, lit all the lighthouses along it. Our method is the universal method for
searching for the strongest chess move, worked out for all positions, without exception. All of them!
Chapter 2
Medium-Difficulty Examples
13...Nxf4
Sacrificing the knight, while at the same time openly assaulting the white bishop! Why? Why, you ask, is Bareev hitting
the bishop, precisely? Where’s the king? Where are the queen and rook?
It’s all pretty simple: Bareev attacks the bishop because he has nothing better to attack.
14.Nxf6+
A mistake. Topalov, it is almost obvious, simply did not notice Black’s shattering move 16.
Better 14.gxf4, although then too, after 14...Bh4+ 15.Kf1 f5 16.Qe2 Kh8 17.Kg2 fxe4, Black has the upper hand
(variation provided by Fritz).
In three moves, the position has changed drastically. And those changes are clearly not in White’s favor , since his
king’s position has gone sharply downhill.
Let me show you: compare the mobility of White’s king after 13.a3 and then after 16.Qe4. Back then we had six
possible moves (not forgetting the right to castle long), and now only two. That means that the probability of a radical
resolution of the position has grown. Look for a sacrifice!
16...Re8
There’s the refutation of White’s plan. Black sacrifices a rook – not for the heck of it, but in order to speed up the attack.
The alternative is the almost prosaic knight check 16...Nd3+. See whether this check wins. And don’t forget to use the
computer’s advice.
17.Qxe8 Bf5
The apex of Black’s grand plan! By sacrificing his other rook, Bareev forces mate.
18.Qxa8
If 18.Qa4, then 18...Nd3+ 19.Kf1 Kxh7 (for example) with an easy win (Bareev).
18...Qe4+
The local density of the white pieces around the white king is exactly: none!
19.Kf2
23...Nb6+
and Topalov resigned, as he cannot prevent mate within a few moves. 0-1
14.g4
Why does this pawn rush forward? For the sake of the seventeenth move!
17.Kg2
White not only “elevates” the king, but also significantly increases the density of packing of his own position, from 9/40
to 9/32. And it’s also important that now White becomes more justified in making an out-of-order exchange.
17...h6
What to do? Draw a “picture”: a2-a3, Ra2, and Rfa1. And then play a3xb4, etc.
18.a3
“A purely strategic error: on 20...Rab8! 21.axb4 axb4, White has the open file, but no invasion squares. And after the
exchange of rooks the white queen will feel like a pirate on the open seas in the enemy’s rear” (Spassky).
And how does that sound, in our terms? 9/40 < 9/32, from which comes 20...Rab8!.
24.Qa2
Spassky continues to play strategically. The queen moves to a2 and later on (see move 29) to b8. But was he correct?
The silicon monster suggests that White had a very strong, concrete move – 24.g5, and then 24...Nxg5 25.Qa2 Bb7
26.Qa7, a strictly “Tal-like” open attack on the poor light-squared bishop. What to do?
Rybka advises: 26...Nxf3 27.Kxf3 Qe7 28.Qb8 Kh7. The evaluation is ± at a depth of 15 ply. So there!
So, why “Tal,” and not “Capablanca”? Because t(23...Bxa8) = 32/23 = ~1.39. Thanks to Black’s mistake on move 20,
we have moved into the Tal Zone and have the right to a tactical strike.
24...Bb7
25.Qa7
Of course, 25.g5 was stronger – see the note to White’s previous move. More than likely, Spassky simply missed this
tactical possibility.
25...Qe7 26.Be2
26...Nhf6
The tenth World Champion thinks that, “... he would have held out longer with a passive approach to defense: 26...Bf8!?
27.f4 f6, with the idea of playing the maneuver 28...Bc8 29.Qb6 Qb7, ridding himself of the white queen’s pressure.”
How is this move notable? By the fact that from the b8 square, the queen keeps watch on, not one, but three targets – the
light-squared bishop, the knight, and the d6-pawn.
It’s easy to see that White has already safely managed to put all his pieces on their best squares. In other words,
Spassky, playing strategically, has fulfilled Point 1 of the Capablanca Algorithm. Spassky – we can say this – has
forgotten “Tal.” For now...
32.h4
With this move, Black chooses the most radical measure to prevent the possible pawn sacrifice 35.e5 fxe5 36.Ne4, and
so on (Spassky’s recommendation). I would add that the maneuver Bf3-e4-f5, followed by Ng3-e4, also looks very
promising.
But it’s not beyond the realm of possibility that, instead of the text, it nevertheless would have been stronger to play
34...Ne7.
Your are hereby tasked with trying out Spassky’s recommendation. Research the position after 34...Ne7 35.e5 fxe5
36.Ne4 – first without a chess engine, and then with one.
35.Be3
The best square for the bishop.
35...Ne7
36.Nxe5 fxe5
37.Bxc5
Clearly, with both queens still on the board, the best square for the white queen is on b8. With his move 39, Spassky
starts a new cycle of the strategic algorithm – the first point of it. But the immediate queen trade (39.Qf5+) deserved
the most serious attention. Why?
Because 7 > 5 and 7/28 > 5/28 – compare the density of packing of White and Black with respect to the king and pawns.
White is more compact, and consequently, is entitled to exchange first!
And now, a word from the unemotional Fritz: 39.Qf5+ Qxf5 40.Nxf5 Ngf6 41.e5 Nd7 42.e6. Is White winning? Please
extend this.
...A few years after Fritz’s revelations, I returned to the position after 39.Qf5+ and fired up Rybka. And here’s what it
told me:
After 39.Qf5+ Qxf5 40.Nxf5, we must play, not 40...Ngf6, but 40...Bc8. And, lo and behold! – after 41.e5 Bxf5
42.gxf5, Black gives back the piece with 42...Ngf6, and there’s no win in sight!
The unanswered questions are: could it be that White should not play 41.e5? And what then? Or, finally, was Spassky’s
39.Qb8 the strongest? Or did Spassky, forgetting about “Tal” (see my notes to 24.Qa2 and 25.Qa7), play on until he
lost real winning chances?
39...Nef6
A mistake. 39...Ngf6 was stronger, and if 40.e5, then 40...Nxd5 with unclear play. And on 40.Nf5, Black has two
replies (40...Qc8 and 40...Qc7) when White is better, but there is no direct win.
40.Nf5 Ne7
41.Nxh6
An elegant little combination in the style of Capablanca. On 41...Kxh6, White wins at once with 42.Qf8+ Kh7 43.Qf7+.
On the board, it’s “Tal.” Why?
I will answer that question with a question: after 40...Ne7, what is the mobility of the black king?
41...Nexd5 42.cxd5 Kxh6 43.Qf8+ Qg7 44.Qxc5 Nd7 45.Qd6+ Kh7 46.e5
Spassky’s outstanding! A final little combination: on 46...Nxe5 the bishop check wins – 47.Be4+.
49.Kg3
17.Kf2
With this move, White considerably increases the density of packing of his king and pawns from 8/28 to 8/21. White
becomes more compact than Black, and that’s a plus! No less important is the fact that after 17.Kf2 the white pieces’
mobility is also increased. Verify this!
However, not everything is so remarkable. The downside is that White has brought his king closer to the front. His
position has become less safe, which means that, in the not-so-distant future, Botvinnik will have to take measures
(which cannot be put off) to increase the safety of his position.
What sorts of measures, exactly? Botvinnik answers (I’m getting a little ahead of myself here). Here’s his deployment:
the bishop on f3, queen on d2, and Nd4-c2-e3. This setup will guarantee White a safe life. Only once it’s completed can
White think about more than defense.
17...Qf6
18.Bf3
18...Ne7 19.Qd2
We know what to do: we increase the density of the pieces around our king.
19...b6
Very passive; 19...0-0 was better. Black need not fear the ghost of an attack. The strong player is always a fearless
player!
20.g4
Given the smallest opportunity, Botvinnik shows some activity. 21.g5 is threatened, and Black cannot reply 21...hxg5.
This aggressive pawn move is a worthy reply to his opponent’s passive stance, and this kind of play earns our
appreciation.
20...g6
Black’s lack of activity is simply revolting! With such a policy, he is doomed to a bad end.
Fritz advises 20...g5 or even 20...0-0. A couple of variations:
a) 20...g5 21.fxg5 (21.g3!?) 21...hxg5 22.Rxh8+ Qxh8 23.Qxg5 Ng6;
b) 20...0-0 21.Rh5 Rfe8 22.Rdh1 Kf8.
21.Rh2
Much too late, and without great hope of success. Black is in a bad way.
24.Ne3
“Here the knight protects the pawns at c4 and g4, and also attacks the points d5 and f5. White has no fear of 24...gxf4
because of 25.Nf5 Bxf5 26.exf5, and then Qxf4, with a triple attack on the pawn at h6” (Botvinnik).
24...Be6 25.g3
With this move, White has completed the setup we discussed in our notes to 17.Kf2. True, Botvinnik strengthened it
with two rooks on the h-file and the pawn at g4. The reason: Black’s passive play.
25...a5
26.a3
Much as a droplet of water is like the ocean, this little move by the white pawn reflects all the primeval depth of our
impressive game. Botvinnik improves this pawn, most probably by intuition, in the most modest way imaginable. Or, to
put it another way, he improves it proceeding out of esthetic and practical considerations. Here’s what the first Soviet
World Champion writes in his annotations: “When your opponent doesn’t have even a hint of possible counterplay,
waiting moves are always allowable and even useful, in order to let the other side express himself, while you yourself
get the time to fully analyze the position again.”
In the language of the Capablanca Algorithm this sounds standard: it’s the second point of the algorithm! Botvinnik
improves the pawn because he has already placed his pieces on their best squares.
The final mistake. 29...Bd7, not allowing the activation of White’s bishop, was better.
33.Nf5
Open attack on the queen! And it is only because the black king is (thus far) out of reach.
Getting to the king. There is no defense against the check from e6.
“Or 36...Qxg5 37.Rh5 Qg7 38.g5 c5 (38...fxg5 39.Be6+ Kh8 40.Rxh6) 39.Be6+ Kh8 40.gxh6 Qxg3+ 41.Kxg3 cxd4
42.cxd4 Nc6 43.R5h4, etc.” (Botvinnik)
37.Be6+ Rf7
15.fxe5
White, without any particular need to do so, exposes his king – the f4-pawn disappears.
15...Ndxe5
16.dxe5
The final mistake. It is obvious that the American grandmaster has overlooked his opponent’s stunning reply.
16...Bh4+
Now 17.Nxh4 is not possible because of the mortal blow 17...Qf2#. The only other move was –
17.Kxh4
– but then
17...Rxf3
18.Rf1 Qb4+ 19.Bf4 Qe7+ 20.Bg5 Qe6 21.Bf5 Rxf5 22.Nf4 Qxe5 23.Qg4 Rf7 24.Qh5 Ne7 25.g4 Ng6+ 26.Kg3 Bd7
27.Rae1 Qd6 28.Bh6 Raf8 0-1
3) a barely noticeable “minus” in the safety factor, because the packing density of the white pieces around the king is
smaller than the corresponding density of the black pieces.
Our preliminary diagnosis: the TC Algorithm, or the right side of “Capablanca.”
For now, we cannot evaluate, with the needed accuracy, the level of safety of the opposing kings. Evidently, we need
more information:
4) the fourth chess factor is at parity;
5) the fifth factor: White stands better here. This is visible to the unaided eye: compare White’s four pawns on the fourth
rank with the sole black pawn pushed to its fourth rank.
White has something to lose (∆(10...c5) > 0), which means that he has something to protect. In other words, the factor of
spatial expansion works against any modifications of the Tal Algorithm. It moves us in the direction of the Petrosian
Algorithm.
Our final conclusion: not the TC Algorithm, but the right half of the Capablanca Algorithm.
How did the game go?
11.dxc5
11...Nxc5
After 11...dxc5 12.e5 (… 13. 0-0-0), White stands clearly better.
But now play becomes forcing.
Black’s last move is no better or worse than other choices, such as 17...Kf8, 17...f5, etc.
Trades favor White, and so this move cannot be bad. However, most likely the requirements of the position are better
satisfied by a different trade: 18.Bxc5. After 18...dxc5 19.Rb1 Rd8+ 20.Ke3 Rd6 (Rybka’s first line), White has every
prospect of winning: 21.Rb5 Rc6 22.Ra5 a6 23.e5 … Ke4-d5, and then Ra3-b3, a4-a5, etc.
18...Ne6
Wrong! Following 18...Rc8 … ...Rc6, ...a7-a6, ...dxc5 (after trading bishop for knight) and ...f7-f5, Black gains real
drawing chances. Why?
Short answer: the fortress!
Of course, White would have every chance to avoid such an unfavorable development, but... but somehow, a direct win
doesn’t seem possible.
Investigate this position! First on your own, and then with the help of a chess engine.
19.exd6
19.f5 was stronger (Fritz). Now 19...gxf5 would be bad: after 20.Rg1+ Kh8 (20...Kf8 21.Bh6+ Ke8 22.Rg8++-)
21.exd6 exd6 22.Bd4+ Nxd4 23.cxd4 … Kc3 and c4-c5, White wins.
Simplest. White improves his pawn without delay. The goal: the a8 square!
In order to increase his ∆(move), White doesn’t shrink even from sacrifices. First a pawn...
25...Nd5
26.Kd3
30...Kd7 31.a7
Black resigned. 1-0
In the final position, ∆(31.a7) = ~2.17. Confirm this.
Flawless play in strategically superior positions correlates with an increasing ∆(move) – that is, with expansion.
3) a significantly higher level of safety, because h7, located in the immediate vicinity of the black king, is attacked by
the white rook.
Our situation is relatively simple. It arises from the great number of examples we have already examined, where the
attack on the king had to be prepared carefully.
14.Ke2
By improving the king, White not only increases his ∆(move) and compactness: his mobility is also enhanced –
primarily, that of the rooks.
White plans to double rooks on the h-file. His goal will be the material target on h7.
16...f5
This move cannot be called adequate. It seems to me that Black has missed the point of White’s operations: White will
have rooks on the h-file, a pawn on g3, and the king on g2. Next he will redeploy the c3-knight to f4. And all this with
that pawn on g5, so awful for Black.
Fritz advises 16...f6.
17.Rh4
17...Rf7 18.Rch1
Open attack on the h7-pawn. By attacking it, we also attack the king. Why?
Because the h7-square is one of the five squares surrounding the king on g8.
18...Bh8
In total accordance with the second point of the Tal Algorithm, White places his pieces on squares suitable for the
coming attack. The knight goes to f4, closer to the enemy monarch!
20...Ne7 21.Nf4 c5
It’s very hard to call this counterplay timely. It looks like Black’s predicament is very bad indeed.
22.g3
White is in no hurry – perfect strategy! Now the king has a very safe and comfortable refuge on g2.
24.Rh6
Fritz likes this move a lot. It also likes 24.b4 and 24.Ne5. Computer programs always think concretely.
24...Rc6
Losing quickly; Black clearly underestimated the flaws in his position following the knight’s jump to e5. It was
necessary to address this concrete threat: 24...Qc6 (Rybka). But even then, after, let’s say, 25.Qe2 Qd6 26.Nh4 Rg7
27.dxc5, White has a clear advantage. More to the point: White’s position is strategically winning!
25.Ne5
Preparations for the offensive are complete. More precisely: the stage is set for the dynamic phase of the attack on
material targets. This knight move is the first shot: a double attack!
26...Qd8 holds out longer, but there too, Black’s predicament is not to be envied – 27.Rd1. One variation runs 27...c4
28.Qd4 b5 29.e6 Rg7 30.Qf6 Qe8 31.Rdh1 Rc7 32.Qxg7+ Kxg7 33.Rxh7+ Kf8 34.R1h6 … Rh8++-.
27.e6
27...Rg7 28.Nxd5
28...Rxe6 29.Nc7
The rook hangs, but worse than that White also threatens 30.Qd8+ Kf7 31.Qe8#.
Black resigned. 1-0
Postscript to this game – the parameters of two positions.
I begin with the position after 13...g6: m = 1, t = 48/32 = 1.50 > tcr = 1.25, “+”, ∆k = 8/28 – 8/32 > 0, ∆(13...g6) =
~0.31.
And now the position after 24...Rc6: m = 1, t = 52/26 = 2.00 >> tcr = 1.25, “+”, ∆k = 8/28 – 8/32 > 0, ∆(24...Rc6) =
~0.54.
Note the significant growth in the “t” parameter, and the insignificant increase in ∆(move) – to be exact, ∆(24...Rc6) –
∆(13...g6) = ~0.23.
And one more thing: White has one rook and one minor piece more than Black in the f1-f8-h8-h1 rectangle.
Solve this puzzle: 0.50 + 0.23 + rook + minor piece = ?
3) a microscopic advantage in the safety factor (the black king’s freedom is somewhat limited, but White lacks the
power to carry out a successful attack on it (we are, after all, deep in the ending).
Our final diagnosis: the position requires the Tal Algorithm. Exactly that, because our t > tcr = 1.25 and because you can
see our (admittedly small, but nevertheless real) “plus” in the factor of safety. We are obliged to attack!
45.h5
After 46.h6 gxh6 47.Rxh6, White intends to win the h-pawn, and with it the game.
But that’s not all. Also very important to us is the fact that once the h-pawn gets traded for the g7-pawn, there will be a
dramatic, favorable alteration in ∆k. Please confirm this on your own! Grim prospects – for Black, of course!
47.h6 g6
Or 47...gxh6 48.Rxh6 b5 49.Rxh7 Ke7 50.Rh6 b4 51.Rb6 b3 52.f6+ Kd7 53.Kg4 b2 54.Kf5 Kc7 55.Rb5 Kd6 56.Rb7
[Zugzwang! – A.Sh.] 56...Rf1 57.Rb6+! Kc5 58.Rxb2+-(from Karpov’s My 300 Best Games).
And if 51...f6 (instead of 51...b3), then White wins by marching over to the b-pawn: 52.Kf2 b3 53.Ke2 b2 54.Kd3.After
54...Rd1+ 55.Kc2 Rf1 56.Rxb2 Rxf3 57.Rb3, White will elevate his king with no particular difficulties and escort the d-
pawn to promotion.
48.fxg6 hxg6
48...Rg1+ 49.Kf2 Rxg6 is also bad: after 50.Rxg6 fxg6, the pawn ending is hopeless for Black. Check it out – don’t use
a computer!
And if 48...fxg6, then White wins by 49.Kh4 (Rybka’s recommendation). White threatens a victorious king march to g7,
therefore 49...Rb3 forcing 50.Kg4 etc., for example 50...b5 51.Rb6 Kc7 52.Rf6 Rb2 53.Rf7+ Kb6 54.Rxh7 b4 55.Rh8
Kb7 56.h7 Rh2 57.d6 and wins.
Curiously, Karpov’s recommendation of 49.Re6 (after 48..fxg6) probably doesn’t win. According to Karpov, 49...b5
50.Rxe5 b4 51.Re6 is winning. However (pay attention now!), there is 51...b3 52.Rb6 g5 53.e5 b2 54.Kg2 Kc7, or
52.Kf4 g5+ 53.Ke5 Rf1 54.Rb6 Rxf3 55.d6 Rd3... draw?! Play it out! On your own – no computer!
49.Rc2
Or 49...Ra1 50.h7 Ra8 51.Ra2 Rh8 52.Ra7+ Kd6 53.Rxf7+- (as indicated by Fritz and supported by Rybka).
50.Rg2
50.Kh2 is even simpler. It’s useless to argue with both Fritz and Rybka: 50...Ra1 51.Rc8+-, or 50...Rg5 51.Kh1+-. It’s
not hard even for a world champion to miss that last “anti-positional” king maneuver to h1.
Computer: you destroy hundreds, thousands of illusions! You force us to think more concretely.
Fortunately, 50.Rg2 doesn’t throw away the win. Karpov has merely delayed the conclusion for two or three moves.
50...Ra1
51.Rh2
51...Ra8
52.Kg4 f6
Otherwise there comes 53.Kg5 … Kf6-g7 or h6-h7 and Kh6-g7, winning in short order.
53.Rb2
53...Kc7 54.Rc2+
Check – that is, an open attack on the king. The apex of the Tal Algorithm!
54...Kb7
55.d6
The white pawn queens after 56...b4 57.Rc8 (see the diagram below) 57...Rxc8 58.dxc8Q+ Kxc8 59.h7 b3 60.h8Q+.
Black is behind by “only” two tempi.
Position after 57.Rc8 (analysis)
Postscript: the dynamic of the fourth and fifth parameters. Let’s compare the positions after 44...Kd7 and 57.Rc8:
∆k = 6/20 – 6/21 = ~0.0 and ∆(44...Kd7) = ~1.29;
∆k = 5/25 – 5/24 = ~0.0 and ∆(57.Rc8) = ~2.33.
With an unchanging ∆k, White substantially increased the ∆(move) parameter. In other words, White won because he
successfully shifted his position’s center of gravity. The direction of the expansion was the queening square!
“Tal” and “Capablanca” are twin brothers, are they not?
3) rough parity in the safety factor (the dark-squared bishop’s pressure on f2 can be disregarded).
Preliminary diagnosis: the Petrosian Algorithm.
4) The fourth factor of the position: parity,
5) The fifth factor: approximate parity, or more precisely ∆(15...Rd8) = ~-0.08.
Since we can ignore the effect of the fourth and fifth factors, our final conclusion strongly confirms our preliminary one:
this position demands the Petrosian Algorithm.
16.Be3
Why?
Because exchanges are very much in our favor, and “Petrosian” requires these exchanges.
With the text move, White significantly increases his army’s mobility. He increases, first of all, the mobility of the dark-
squared bishop itself, but also that of both rooks.
16...Kf8
Of course, 16...Bxe3 is bad for Black, because after 17.fxe3 the f1-rook’s mobility increases sharply. It’s also important
that, after the trade on e3, White’s king has the possibility of immediate elevation to f2. This would significantly
increase the density of his position, which is a good thing for White.
Other candidate moves: 16...Bb8 and 16...Nf5. Investigate whether they work out concretely. First without the
computer’s advice.
20.Nc5
20...Nf5
“Nor would 20...Ng6 21.Bf3 Rc7 22.b4 have improved matters; and after 20...Rc7 21.Bd3! Black’s ideas of transferring
the knight to d4 would have been reduced to an impossible dream. But he ought to have preferred 20...a5, hoping for a
little counterplay in the variations 21.Bf3 a4 22.Rd6 Ba6, or 21.Kf1 a4 22.Bf3 Ra5 23.Ne4 Be6” (A. Kuzmin).
The following line, which demonstrates the viability of Black’s position, is curious: 20...a5 21.Rd6 a4 22.Bd1 a3 23.b4
Ra8 24.Ba4 Bf5 (… 25...Rb1) 25.b5 Rd8. Black’s last move was indicated by Fritz. Now Black has realistic drawing
chances!
21.Bg4
Of course! It’s the only way! Now the black knight is paralyzed.
21...Ke7
Covering the d6 square, because White threatened to take on f5 and get into d6 with the rook.
22.Re1
22...Kd6 23.Ne4+
Once more, the Tal Algorithm.
23...Kc7 24.c5
24...f6
A mistake. Black most likely overlooked his opponent’s powerful reply. Better (at least according to Fritz) is 24...Ne7
25.Bxc8 Nxc8 26.Ng5 f6 27.Nxh7 Rb7 or 27...Ne7, with vague hopes of salvation.
25.f4
The decisive blow! Now 25...exf4 loses immediately, as after 26.Nd6 Black is defenseless. Verify this, but without
resorting to a computer.
“The last chance here might have been to transpose into a knight ending after 28...Kd8 29.Rxe5 Re7” (A. Kuzmin).
Or 31...Nxc5 32.Nb6+ Kd6 33.Rd4+ Ke5 34.Rc4 Rb8 35.Rxc5+ Kd6 36.Na4+-.
35.Nd3
36.Nf4+- is threatened.
35...Kf6
If 35...Rf8+ 36.Ke3 axb4, then 37.Nf4 Rxf4 38.Rxf4 Nxf4 39.Kxf4, with a winning pawn ending.
White to move
We are playing White. We have:
1) parity in the first factor;
2) an overwhelming advantage in the time factor: t = 63/33 = ~1.91;
3) an insignificant advantage in the safety factor.
A unanimous diagnosis: the Tal Algorithm.
1.Qxg7+
An open attack on the king + a sacrifice! Of course, it needs to be calculated exactly and all the way to the end.
Or 5...Kxh4 6.Rh6#.
6.f4+ Kxh4 7.Kh2 1-0
White to move
We play White. We have:
1) parity in the first factor;
2) the Tal Zone: t = 43/32 = ~1.34 > tcr = 1.25;
1.Qxd7
Magical play!
6...Bxg3+ 7.Kg2
Black resigned. The only way to avoid checkmate is 7...Bh2 8.Rh1 Qg4+, etc. It avoids mate, yes – but not ultimate
defeat. A treasure! 1-0
28.Na5
28...Bc8 29.Kf2
Pawns are chessmen too, and they take part in White’s intended deployment (see the position after 36...Ne7).
31...f6 32.Bc2
The place for the light-squared bishop is on b3. From here it will fire on the newly created target at e6.
32...Ndb6 33.Bd4
From this square, White’s bishop not only hits the b6-knight, it also paralyzes the target on a4.
White steadily increases his ∆(move). The packing density of his king and pawns exceeds the opponent’s corresponding
indicator twice over.
Here, I note merely that ∆(36...Ne7) = 1.10. Compare this with ∆(27...a6)!
37.g5
The first point of the strategic algorithm is behind us (put your pieces on their best squares!). And the second one is also
behind us (advance your pawns!). I hope I don’t need to repeat again the already registered truth of the third point of
the algorithm:
White plays for exchanges!
37...fxg5
42...Kxe7 43.e5
A new setup (see also 45.Ke5).
46.Bd1
46...Be8
Or 46...Bc6 47.Bc2 Be8 (47...Bd7 48.Bb3) 48.Be4 Bd7 49.Bg2 Be8 50.Bf3 Bd7, and we reach the same position as
after 47...Bd7 in the game.
Black will be down a pawn after 51...Bd7 52.axb5 axb5 53.Bc2 and 54.Bd3. And if 51...bxa4, then 52.Bxa4 Bb7
53.Bb3 Bc8 54.Bc4, and we have a new Zugzwang on the board.
In this position, (54.Bc4) = ~2.17. And now, dear reader, compare the center of gravity of the white and black forces
after 54.Bc4, 45...Kc7, 36...Ne7, and 27...a6.
Done? I get 2.17, 1.43, 1.10, and 0.00.
At all stages in this battle, the density of packing of White’s position was never lower than that of Black’s position.
White’s strategy was flawless!
26.Qg3
Because ∆k = 0, White has the right to offer a preliminary exchange. However, we could also put it like this: with this
move, White hits the f6-pawn with the threat of 27.Qxh4 Nxh4 28.Rxf6.
26...Qg5 27.Kh2
With this move, Sasikiran considerably increases the compactness of his position. Now his king and all five of his
pawns are contained in the rectangle a2-a4-h4-h2, and not a1-a4-h4-h1. White has become more compact than Black!
27...Qxg3+
One cannot call Black’s play consistent, let alone exemplary. If the queens were to be traded, then this should have
happened one move earlier, without the loss of tempo. And if they were not, then 27...Kg7, elevating the king, was
more consistent!
28.Nxg3 Rf8
Something momentous has occurred: the two strongest pieces are off the board. Play has simplified, and the safety
factor has taken a back seat...
Here’s some very useful and timely practical advice. After a queen trade, try to forget at once the previous diagnosis of
the position. Analyze it again, look at it with fresh eyes!
Now we have m = 1, t = 32/26 = ~1.23, “=”, ∆k > 0, and ∆(28...Rf8) = ~0.11.
Our new diagnosis: the TC Algorithm, as the new starting point lies somewhere on the border between the algorithms of
Capablanca and Tal. We are entitled to play more aggressively!
Here’s some very useful and timely practical advice: After a queen trade, try to forget at once the previous
diagnosis of the position.
Analyze it again, look at it with fresh eyes!
29.Nh5
White hits the material target on f6. This is “Tal,” but it’s also “Capablanca,” since White improves the knight. All
together – harmony!
29...f5 30.Kg3
We elevate the king. The direction of the expansion is the eighth rank.
30...Ne7
31.Rc1
The best square for the rook. White threatens the close-to-fatal 32.Rc7.
31...Rc8 32.Rxc8+
32...Bxc8 33.Kf4
33...Kf7
34.Ke5
Flawless, both strategically and tactically! White has calculated everything out perfectly. He is playing “Tal”-like, for
the destruction of his opponent’s spirit.
34...Nc6+
35.Kxd5 Bb7
Threatening 36...Nb4+ 37.Kc4 Nxd3 38.Kxd3 Bxg2, when Black puts the worst behind him. What do we do?
36.Nf4
That’s what! From f4, the knight securely holds both the bishop and the g2-pawn.
White retreats because he has something to defend – an extra pawn. He needs to think about consolidating his very
scattered position; and Sasikiran, as we can see, plays to increase his pieces’ packing density. Bravo!
36...Kf6
Threatening 37...Ne5+.
37.Kc4
37...Kg5
He could have prolonged the struggle with the solid 37...Ne7 … ...Nc8-d6 (Rybka).
38.g3 h5 39.d5
The beginning of the end – White’s passed pawn has crossed the midpoint line.
Kuzmin writes: “On 41...Kf6, of course, Sasikiran was planning, not 42.gxh4? because of 42...Nb3+ 43.Kc4 Na5+,
when he would have had to give up the d5-pawn, but 42.Bc2!, threatening to capture the knight. And if 42...Ba6, then
43.Nf4 was very strong.” After 43...hxg3 44.h4, White must win.
43...Kf3 44.Nc1
46...f4 47.Ne2
14...e4
We attack the queen, and only because we cannot get at the king – for now!
15.Nxe4
Only heavy computer analysis can answer the question: which is better, 15.Nxe4 or 15.Qc2?
17.Qc2 Nd4
A second open attack on the queen. For now, everything looks (I’m thinking of Ivanchuk’s attack) almost trivial...
18.Qb2
18...Nxe2
Brilliant! However, we do not have the right to call this move the key of the idea. The key to Black’s truly magnificent
plan lies in the quiet move 19.
19.Kxe2 Rfe8
That’s it!
20.Qb4
No rush!
24.Qc4+
Spite check.
24...Kh8 25.Re1 Rxe4+
17.Kg2
White improves his king. And not just for the heck of it, but with a plan: he wants to bring both rooks to the h-file,
closer to the enemy king.
17...Qe7
I won’t rule out that it would have been better to open an escape hatch with 17...Re8 and then ...Kg8-f8-e7, ...Bb7-c8,
and, if necessary, ...a7-a6.
Now Black’s plan becomes clear: he intends to put the dark-squared bishop on h6. But there is a hole in this plan.
The stronger side is completely mobilized. Perhaps there needs only to be a knight on b5, but it would stand unsteadily
there: 22.Nb5 a6. If that’s the case, then we must fight for that square! White needs to find a tactical stroke – or its
analog, a strategic exchange.
22.Bd7
On 23...Kf8, Fritz offers an uncomplicated variation: 24.Ng5 Bc8 25.Qf3 Ke7 26.Nxf7, etc.
26.Nb5 Qd8
If 26...a6, then 27.Nh4 is a possibility, and the win is not far off. A reasonable variation (Fritz) runs 27...Kf8 28.Nxc7
Rc8 29.Ne6+ Ke8 30.Nxg6 fxg6 31.Rxg6 Rxg6 32.Qxg6+ Kd7 33.Qf5, etc.
Your task is to extend this variation without the computer’s assistance.
27.Nh4 Nxe4
The American GM Reuben Fine once noted that in such positions, “combinations are as natural as a baby’s smile.” Of
course, he had in mind open positions with overwhelming forces in the attacking sector.
28.Rh7+
29...fxg6 is bad because of 30.Qh6+ Ke8 31.Nxc7+ and wins. Still worse is 29...Rxg6 30.Rh8+ Kg7 31.R1h7#.
32...Qxc7 33.Qxg6
33...Qe7 34.Rxf7 Qxf7 35.Rh8+ Ke7 36.Rh7 Rf8 37.Qe6+ Kd8 38.Rxf7 1-0
3) parity or at best a microscopic advantage in the safety factor (the black king is less than fully mobile, but attacking it
is hardly practical – we are deep into a rook ending)
Our unanimous conclusion: the Tal Algorithm.
“Tal” in the endgame? How can that be?
The answer: “Tal” is “Tal”! And it starts with open and direct attacks on material targets.
37.g4
Not just an open attack on an enemy pawn, but also a sacrifice. White can sacrifice a pawn because his “t” is through the
roof.
The alternative is the quiet, “Capablanca” move 37.Kd3 with the idea of a2-a3, b3-b4, etc.
37...fxg4 38.f5 Rc6
After 38...exf5 39.Rxf6 (39.Kf4 would be bad, because of 39...Rc6 40.Rxa7+ Ke6, and Black has every chance of
drawing) 39...Rc8 40.Rxf5, White restores material equality with an obvious advantage and good winning chances.
39.Rxa7+ Kd6
40.Kf4
40...h5 41.Rf7
41...Ra6 42.Rxf6
He can play this way, too, but Fritz likes 42.a4 better, with this variation: 42...Rb6 43.Rxf6 Rxb3 44.Rxe6+ Kd7 45.a5,
etc.
45...Rb1
46.Kh4
15.Qg5
A worthy reply!
White is entitled to an out-of-order exchange (“Capablanca”), and he is required to make use of it (“Petrosian”)!
Outstanding strategy. Tactics? Can Black stop to munch on the pawn at f2?
The unanimous answer is: No! 15...Qxf2 loses by force: 16.Rf3 Qg1 17.Qg3 … 18.Be2+-. Nor does 16...f6 save him:
after 17.Qg4 f5 18.Qh3, the queen is captured by force.
Black must acquiesce to the queen trade. The game continued
and White got a clearly better endgame. Spassky won on move 31.
39.Kg5
39...Qxg2+ 40.Kh6
The king is safe here, securely protected by the enemy (!) pawns on g6 and h5. Now White has only to put his queen on
the long dark-squared diagonal.
40...Qb2
Clever, but alas it is not enough. Black intends to parry the threat of 41.Qf6 by hopping the knight to e6.
41.f4
All of White’s pieces are placed ideally. Isn’t it time to elevate that pawn?
The Capablanca Algorithm? Yes or no?
Black resigned, as there is no defense against 42.Qe5, for example 41...Qc3 42.Qe5 Ne6 43.Nxe6 Qxe5 44.fxe5 fxe6
45.Kxg6, etc. 1-0
Why? Because White has clearly superior forces in the attacking sector. This means that the likelihood of blood being
spilled increases sharply. Timman’s double sacrifice has good chances of being correct.
For us, it’s very important that this double exchange sacrifice is an intuitive one. It’s practically impossible to calculate
its consequences all the way to the end. This is a task that even the most powerful computer program would find hard to
deal with. It remains but for us to go by feel, from one favorable position to the next – all the way to victory!
18.Qf2
White’s intention is primitive enough: 19.Qxf6 … 20.Bd4 and Qg7#. Other possible attacking continuations: 18.Bxh6,
18.Bd4... How many others?
Follow up on these continuations – first without a computer, then with one.
In the diagram position, if Black now plays 18...Qe7 (… ...Rfe8 and ...Qf8 – closer to the king!), then, according to
analysis by V. Atlas, 19.Qh4 looks very strong: 19...Rfe8 20.Bd4 Kh7 21.Bxf6 Qf8 22.e5 d5 23.Bf1. Does White win?
Yes or no?
Most likely not: after 23...Qc5+ 24.Kg2 Black has the fantastic answer 24...Bf5! (Rybka) when, after 25.gxf5 Qe3, or
25.g5 Qf8, White has nothing decisive – dynamic equality! It’s like a miracle, isn’t it?
Stronger than 24.Kg2 are the alternatives 24.Kh1 and 24.Kh2. Followed by variations upon variations... Thirteen and
even more ply deep, Rybka takes the position to be “²”, but nothing more than that. A flesh-and-blood chessplayer
cannot go that deep!
It is indeed possible that the strongest move in the position is 18...Qa5 (Van Wely’s recommendation). Is that right?
I admit to thinking long and hard over this, and I thought it was not so. My skepticism was fueled by Fritz – at that time
Fritz 8, back then the latest version of that miracle program. We managed (more Fritz than I) to find a very dangerous
variation for Black: 18...Qa5 19.Bxh6 Qe5 20.Bf4 Qc5 21.Be3 Qe5 22.h4, depriving the queen of the g5 square and
preparing the murderous 23.Bd4 … 24.Qxf6.
But, in the age of Rybka, a bitter disappointment lay in store for me. In the first place, Rybka holds the position after
22.h4 with 22...Qa5 23.Bd4 Rfd8 24.Qxf6 Kf8 25.h5 Ke8 … 26...Kd7. Black is in a bad way, but it’s ± and not +- . Of
course, by digging deeper into the position, I might turn that ± into a +-. I didn’t do this. Why not?
Because – and this is in the second place – my silicon friend found an unexpected reply to 19.Bxh6 – 19...Qd8!, with
the continuation 20.Be3 Kg7 21.Bd4 Rg8 22.g5 Kf8 23.Bxf6 Qa5 24.h4 Rc8, etc. What the heck?
Could it be that the position after 24...Rc8 is one of dynamic balance? Or is White only a bit better after all? Or...?
Chess is inexhaustible, and the truth – which we equate with the strongest move – drifts away from us. It loses itself in
the wasteland of the primeval chessic chaos, and chaos in chess is irrational complications.
Van Wely preferred to make a king move:
18...Kg7
Black defends the targets at f6 and h6, and also sharply increases the local density of his pieces right around the king.
But there is a drawback to this: the king moves closer to the front line, and that’s dangerous – very dangerous!
19.e5
Powerful, and very strong. The threat is to win the f6-pawn with check.
19...fxe5
Most likely, Van Wely overlooked his opponent’s spectacular reply. On the other hand, Black no longer has a defense.
At least, neither Fritz nor Rybka itself could see one. Here are some variations:
19...Rh8 20.Qxf6+ Kg8 21.exd6,
a) 21...Qxd6 22.Bd4 Rh7 23.Ne4 Qc7 24.Qf2+-;
b) 21...Qd8 22.Qe5 Rh7 23.Ne4, etc.
20.Bxh6+ Kg6
21.Qh4
Black resigned in view of 21...Rg8 22.Be4+ f5 23.Bxf5+ Kf7 24.Qh5+ Ke7 25.Bg5+. No comment necessary. 1-0
GM Psakhis is almost pacifistic. He patiently prepares exchanges – White intends to play 21.Nc5. Exchanges favor
White, but not all of them do. For some reason, Psakhis isn’t satisfied with the variation 20.Nc5 Bxc5 21.Rxc5 Rxb2.
Why?
The answer is: compactness!
20.b3 Qd5
A good move – Black prevents (if only for a while) the knight sortie to c5, which would be very unpleasant for him.
An alternative to this careful move would be the aggressive 20...g5, raising his sword and seeking to test the soundness
of White’s citadel. After 21.Nf3 h4 and then ...hxg3 and ...Qh3 he would obtain real counterchances.
What should White do after 20...g5? Answer blow for blow! Like 21.Nf5!. For this we have a sound basis. Only by
force, only with “Tal,” can we bridle our opponent’s rage. And no doubts. We must win!
Now the straightforward 21...Qxf5 loses – 22.Rxc6 decides. Check it out!
And if so, then Black has only one reasonable reply: 21...Rg6. Black delays capturing that cheeky knight. Continue:
22.f3!. The battle of nerves reaches its climax! Now Black must accept the sacrifice, since 23.g4, with a complete
positional bind and a guaranteed win, is threatened. But the sacrifice cannot be accepted right away – play 22...Rb4 first
(recommended by Fritz).
Then 23.Qxa5 Rb5 is quite possible. Black, of course, would not object to a repetition of moves: 24.Qa4 Rb4 25.Qa5
Rb5, etc. Draw? Answer, White!
Our proud reply is: No!
Let’s examine the position after 24.Qa7 Qxf5 25.Rxc6 Bxc6 26.Rxc6:
21.Qc4
Remember, we are playing by the TC Algorithm, which means that we are entitled to an out-of-order exchange.
21...Kd7
A strategically perfect move. Black “elevates” his king, automatically increasing the density of packing of his king and
pawns.
22.Qxd5+
The queen trade is unavoidable, and therefore it need not have been forced. White erroneously rushes matters. 22.f4 was
better, when if 22...Ba3, then 23.Rf1 with outstanding kingside prospects.
22...Rxd5 23.Nc5+
Careless! It was better to retreat, not the knight, but the rook from c5 to c4. This is because, in the position after
24...Rh8, Black’s superiority in forces on the queenside – the focus of the battle – is plain to see. That place is the
rectangle a8-a1-d1-d8. The black king, rook, and bishop are stronger than the two white rooks, which means that White
simply must surrender this sector to his opponent.
25...Rb8
Returning the favor! Black misses his chance. Instead 25...a4, intending to trade his “poor” a-pawn for the “good” white
b-pawn, was necessary. This would be very good for Black in terms of compactness: after the a-pawn disappeared,
Black would enjoy a significant increase in the packing of his king and pawns.
Now let’s hear from Rybka, fleshing out (and in some cases confirming) Fritz: 25...a4 26.b4 a3 27.Nd2 Rb8 (the target
of the coming attack is the b4-pawn) 28.R5c4 Rb5 29.Ne4 Ke7 30.f4 Ba8.
Dynamic balance! See for yourself!
Finally!
27...Kc7
We have: m = 1, t = 31/30 = ~1.03, approximate parity in the third factor of the position, ∆k > 0, and ∆ (27...Kc7) =
~-0.73. Diagnosis: the TC Algorithm.
“Nothing new under the sun” – compare this diagnosis with that of the position after 19...Rh6. Algorithms for seeking
the strongest move are sometimes very stubborn!
A new question: Why is that? I reply evasively, half in a riddle: fixed pawn chains... Am I right?
White improves his position, increasing the ∆k and ∆(move). It looks strategically flawless!
29...f5
A very suspicious move. Quite probably it’s connected with a tactical oversight – see White’s move 34. Obviously,
Black wants to put his pawn on c5, bringing to life his inactive bishop. But...alas!
33...Rb7
See the note to 29...f5. I believe that my suspicions may have had some basis in fact.
34.Ne4+
The decisive stroke! And it’s no accident. This game-winning check is a consequence of Black’s flawed play. He was
punished for his unjustified ambitions in an inferior position. Black treated his position aggressively (28...Rb4, 29...f5,
32...c5), and this is what destroyed him!
You need to understand: the TC Algorithm for White means the CP Algorithm for Black. Otherwise, you have a game
working against the requirements of the position. Tukmakov was playing almost in “Tal”-style when he should have
been playing almost like “Petrosian,” and that led to this catastrophe.
The TC Algorithm is anti-symmetrical to the CP Algorithm (see the Algorithm Drift Chart), and these two algorithms
are equidistant from the “center” of the Capablanca Algorithm. And the Capablanca Algorithm is self-contained. It is
symmetrical (mirroring) unto itself. On the edges of the spectrum of all attacks and defenses there lie the algorithms of
Tal and Petrosian, mirroring each other.
The rest of the game is of little interest to us – it’s agony for Black. It ended:
34...fxe4 35.dxe4 Rf7 36.exd5 Rxf4+ 37.Ke1 Bxd5 38.Rxc5 Rg4 39.Rxa5 Rg1+ 40.Kd2 Rg2 1-0
14.Rfd1
Wrong rook! In My 60 Memorable Games, Fischer writes: “Correct was 14.Rad1!. Originally I gave the following �?
refutation’: �?14...Ne4 15.Nxe4 dxe4 16.Bxe4 Qxd2 17.Rxd2 Nc4 18.Bxa8 Nxd2 19.Rd1 Nc4 20.bxc4 (best)
20...Rxa8, regaining the pawn with a big endgame advantage.’ But Averbakh found a �?hole’ in my analysis with
20.Bc6! (instead of 20.bxc4 which I had carelessly given as �?best’): 20...Nxa3 21.Bxe8 Bxe2 22.Rd7, and White is
the one who wins instead of Black!.
“I spent an evening just staring at the position after 14.Rad1, trying everything, unwilling to let my brilliancy go down
the drain. The more I looked, the more I liked White’s game!...”
Fischer concludes: “Finally I found 14...Qc8! – the only move to keep the pressure...”
This position attracted notice from a lot of gifted analysts. It was analyzed by Hübner and Kasparov... For example
(speaking of “14...Qc8 – the only move to keep the pressure”), here is what the thirteenth World Champion wrote about
it (My Great Predecessors, Part IV): “According to analysis by Hübner, this is correct both in the event of 15.Nxd5
Nxd5 16.Bxd5 Rd8 17.f4 Rxd5! (17...Ng4!? 18.e4 Nf6) 18.Qxd5 Bb7! 19.Qd2 (19.Qd8+? Qxd8 20.Rxd8+ Rxd8
21.fxe5 Rd2! 22.Rf2 Bxe5) 19...Qc6! (not Fischer’s variation 19...Qh3? 20.Nd4 Ng4 on account of 21.Nf3! h6 22.Bc1)
20.Nd4 Qh1+ 21.Kf2 Qxh2+ 22.Ke1 Qxg3+ 23.Qf2 Qxf2+ 24.Rxf2 Ng4 25.Re2 Ba6 or 20.Qd5 Qxd5 21.Rxd5 Bxd5
22.fxe5 Bxe5 with the better endgame, and after 15.Qc1 Qf5! (Fischer’s recommendation of 15...Ne4 16.Nxd5 Bxe2
17.Bxe4 Kh8 is unclear in view of 18.f4!), but with 15.Bd6! Ne4 (15...Nd3? 16.Nxd5; 15...Qg4!? – G.K.) 16.Nxe4
dxe4 17.Bxe5 (17.Bxe4?! Nc4!) 17...Rxe5 18.Rc1 and 19.Rfd1 White maintains the balance.”
Now I will tear a couple of fragments out of this long game, and fill them out with the ultra-intelligent Rybka.
First fragment: 14...Qc8 15.Nxd5 Nxd5 16.Bxd5 Rd8 17.f4 Rxd5 18.Qxd5 Bb7 19.Qd8+? Qxd8 20.Rxd8+ Rxd8
21.fxe5 Rd2!? 22.Rf2 Bxe5. 22...Rxa2 looks stronger, for instance 23.Bd6 Ra5 24.b4 Ra3 25.b5 Re3µ.
Second fragment: 14...Qc8 15.Bd6 Nd3 16.Nxd5. Contrary to Kasparov, Rybka does not consider White’s last move the
refutation of the “erroneous” 15...Nd3, and continues 16...Nxd5 17.Bxd5 Rd8 18.Bxa8 Qxa8 19.Ba3 Bf6 20.e4 Qxe4
21.Qe3, when White (most likely) does indeed hold the balance. But nothing more than that...
My commentary, and the commentaries of Fischer, Hübner, and Kasparov: if “Tal” (i.e., 15.Nxd5) doesn’t work, then
“Capablanca” must be tried – specifically, the second half of the TC Algorithm; see the diagnosis for the starting
position. That’s where 15.Bd6! comes from. Trading his dark-squared bishop for Black’s knight would favor White.
Why? You know the answer: White is more compact!
After 14.Rad1, Black’s strongest move is probably not 14...Qc8, but 14...Qd7!? (Kasparov’s marks). After that Fischer
advises “15.Qc2 followed by Rd2 and Rfd1 (if 15...Rac8, 16.Qb1!).” However (and here I let the thirteenth World
Champion have the floor), “...after 16...Qg4! 17.h3 (17.f3?! Qh5 leaves the e3-pawn too weak) 17...Qh5 18.Nf4 Nf3+
19.Kh1 Qh6 the initiative is with Black...”
Nevertheless, it seems to me that after 19...Qh6 White should not be worse, most of all because the overall diagnosis of
the position after 13...Bxe5 is quite satisfactory for him. This is why I advise you, dear reader, to extend Kasparov’s
analysis. Concretely: check out 20.Ncxd5.
A question for you: does this extremely curious move work out?
Let’s return to the game, to the position after the careless 14.Rfd1. With this move by “the wrong rook,” White
“catastrophically weakens the f2-pawn” (Kasparov). And he will pay for it!
14...Nd3
With the unpleasant threat of 15...Ne4 16.Nxe4 dxe4, and Black gets an “eternal” knight on d3.
15.Qc2
Not hesitating for a minute, Byrne attacks the enemy knight and – loses by force, since after 14.Rfd1 Nd3 White is not
entitled to attack. His kingside is weakened, and therefore he must not attack, but rather defend. Byrne’s move was his
second consecutive error – a grievous one, and his last!
The strongest move was 15.Nd4 (Kasparov), continuing with 15...Ne4 16.Nxe4 dxe4 17.Bb2 Rc8 18.a4 Qg5 (according
to Fischer, Hübner, and Kasparov). And here “Black has an obvious advantage , thanks to his powerful knight at d3”
(Kasparov).
I add: an advantage, but not a forced win.
15...Nxf2
18...Nxg2
“A brilliant stroke. Byrne was hoping for 18...Nxd1 19.Rxd1 with an unclear game” (Kasparov).
19.Kxg2 d4
20.Nxd4 Bb7+
Check is an open attack on the king, a piece of limitless value. White is finished.
21.Kf1
Kasparov: “The machine confirms that other moves were equally bad: 21.Kg1 Bxd4+ 22.Qxd4 Re1+!, or 21.Kf2 Qd7!
22.Rac1 Qh3 23.Nf3 Bh6 24.Qd3 Be3+ 25.Qxe3 Rxe3 26.Kxe3 Re8+ 27.Kf2 Qf5! (Fischer).”
21...Qd7
White resigned “in view of 22.Ndb5 Qh3+ 23.Kg1 Bh6, or 22.Qf2 Qh3+ 23.Kg1 Re1+!! 24.Rxe1 Bxd4” (Kasparov).
0-1
21...Ng8
The cold-blooded Fritz considers this the strongest. And, true, we find it impossible to argue with it, because chess
engines are head-and-shoulders above humans in sharp positions. And we don’t argue. Even when the move goes
against our judgment.
Dolmatov and Fritz have “neglected” the value scale. They disapprove of the “correct” moves 21...Qc7 (… 22...Qe5+ or
22...Nc2+) and 21...Be2. Both human and computer refrain from open and direct attacks on the king and queen! Why?
Because after the natural
22.Nxg8
22...Qd5
This unexpected move wins at once, as White has no satisfactory defense against 23...Qxa2 or 23...Qxe5+. Cute, yes?
Yemelin resigned after
The value of an empty square on the chessboard can be greater than the value of a square that is occupied by
an enemy pawn or piece. The chess value of an empty square is variable – it can change from the
vanishingly small to the boundlessly great.
Let us recall what happened in the game (and remember: we consider Dolmatov’s play to be ideal):
21...Ng8 was the strongest move – meaning that, in the position preceding this move, the value of the empty square g8
was greater than the value of any other vacant or occupied square available to Black. With this move, Black attacked a
material target on e7 that was putting him in danger. The empty square g8 dovetailed with the occupied square e7...
22...Qd5 was a best move, too. This means that, in the position after 22.Nxg8, the value of the empty square d5 was
greater than that of any other square, including the now non-vacant square g8. With this move, Black attacked targets
on a2 and e5. And also a target on g8, and many vacant squares, whose accurate chess value could not be determined. I
will say only that the empty squares meshed with the non-vacant. A Gordian knot!
23...Qh1+ was a strongest move. That means that the value of the empty square h1 was greater than the value of any
other square on the board.
Vacant squares and occupied squares are interconnected. They make for a unified whole. If we separate them, that
means that we tear them apart. If we reduce chess too much to its constituent elements, we risk killing
chess.
On the other hand, if we do not tear it apart, then we risk turning into complete and utter bystanders in fruitless
contemplation, and along the way, chess kills the chessplayer...
Where does the truth lie?
The truth, slipping away from us, is to be found in that undefinable “middle way.”
3) serious problems with the third factor: we cannot accurately define the level of safety of the kings (fortunately for us,
our king is presently not under any sort of pressure from our opponent, but unfortunately for us it is very loosely
packed; whereas the black king, although not very mobile, is at the same time very securely tucked away).
Clearly, we need more information. What do the fourth and fifth factors tell us?
4) the compactness factor is practically catastrophic! ∆k << 0, since the white king and seven pawns stand within the
rectangle a1-a5-h5-h1 (40 squares), while the black pieces are inside the rectangle a7-a5-h5-h7 (24 squares). This
means that the fourth factor pushes us from the “Tal” to the “Petrosian” side. Perhaps our starting algorithm is
“Capablanca”?
5) the factor of space expansion is virtually at parity. Indeed, ∆(23...Kg7) = -0.10 – that is, very close to zero. In other
words, the fifth factor does not influence the process of determining the algorithm.
Time to add everything up. Weighing all the factors, and laying it all out by parts, I take the risk of pronouncing our
final diagnosis: this position requires an expanded version of the TC Algorithm. I add: our algorithm has a wide
“amplitude” – from “100% Tal” to the quietest “Capablanca.” And yet, the truth eludes us! Why?
Due to the uncertainty in the third factor of the position...
What happened in the game?
24.Qd3
A “Tal”-style move. White not only assails the queen, but also launches a combination. After the very likely 24...Qxd3,
which is not unobjectionable, there would follow, not 25.cxd3, but 25.Re7+ Rf7 26.Rxf7+ Kxf7 27.cxd3:
Position after 27.cxd3 (analysis)
Before us we have a complex and almost pure pawn ending. “Almost,” because for the time being we do lack an
accurate answer to the question: is it a win for White, or a draw, after all? A difficult question!
At the time – in that far-off, pre-computer era – I spent some intense and very happy hours analyzing this ending. And I
well remember how I “succeeded” in finding the win. But then – horrors! – I was shown, convincingly, that my
analysis had a hole in it... The question I posed remains unanswered. Reader, you try!
Here are a few variations, and the questions stemming from each. I think they will help you analyze.
a) The black king goes to the center: 27...Ke7 28.Kg2 Kd6 29.f4 gxf4 30.Kf3 Ke5 31.d4+ Kxd4 32.Kxf4.
Questions:
a1) Does Black have a draw after 32.Kxf4?
a2) Wouldn’t it be better for Black to refrain from 29...gxf4 and allow White to trade on g5 – after 30.fxg5 fxg5, isn’t
there a draw on the board?
a3) Wouldn’t it be stronger to play 29.Kf3 followed by 30.Ke3 … f3-f4 (bringing the pawn to f4 after completely
exhausting his spare tempi with pawns)?
b) Black, wasting no energy, maintains a seventh-rank defense: 27...c5, ...h7-h6, ...Kf7-g7-h7-g7, etc. One variation:
27...c5 28.Kg2 h6 29.Kg3 Kg7.
Questions:
b1) Does Black have a draw after 30.h4 gxh4+ 31.Kxh4 Kh7 32.Kh5 Kg7 33.f3 Kh7 34.f4 Kg7 35.g5?
b2) Does Black have a draw after 30.f4 Kh7 31.fxg5 hxg5 32.h4 gxh4+ 33.Kxh4 Kh6?
b3) Is the penetration of White’s king to b7 dangerous for Black? I.e., 30.f4 Kh7 31.Kf3 Kg7 32.Ke4 Kf7 33.fxg5 hxg5
34.Kd5 Ke7 35.Kc6 Kd8 36.Kb7 (White forces Black to waste one of his few remaining tempi) 36...a5 37.Kc6,
followed by a march to g3 to enforce h2-h4.
b4) Shouldn’t Black, in variation (b3), play, not 32...Kf7, but 32...c6 to prevent the enemy king from reaching b7?
c) Black, not fearing 28.d4/b4, which paralyzes the c6- and c7-pawns, maintains a totally boring seventh-rank defense
by 27...Kg7 … Kf7-g7-f7, etc.
The clock ticks unpleasantly; time trouble draws near. Feverishly you calculate and calculate, looking many
moves deep. You grow weary, and “the taxi’s here – time to go!” You almost despair… What to do??
Believe in your intuition! And, what is most important, banish all fear!
Make your move, without thinking of the consequences.
Step over the precipice… along with your opponent!
Is it a draw, or a win?
We don’t have the right answer. But we do know that the position after 27.cxd3 requires “Capablanca.” And that means
that we must “elevate” this position. When playing to win, we are simply obligated to increase the ∆(move) of our
position!
How do we do this, specifically?
The clock ticks unpleasantly; time trouble draws near. Feverishly you calculate and calculate, looking many moves
deep. You grow weary, and “the taxi’s here – time to go!” You almost despair... What to do??
Believe in your intuition! And, what is most important, banish all fear! Make your move, without thinking of the
consequences. Step over the precipice... along with your opponent!
Chess is a most complicated game. It’s almost inexhaustible, and the element of uncertainty in it was, is, and ever shall
be. The uncertainty of our game is ineradicable. I call upon you, dear reader, not to fear it, but to embrace and to fall in
love with it!
However, let’s return to our starting position, the one after 23...Kg7.
An alternative to White’s aggressive sortie 24.Qd3!!+- (all signs are Adorjan’s) was the modest 24.Qe3. With this
move, White improves the placement of his queen – the first point of the Capablanca Algorithm. This might be
followed by Qe4, Kg2, Re2 – pawns on a2, b3, c4, f2, f5, g4, h3, etc. And “etc.” here means: some exchange or other.
All in complete accordance with the Capablanca Algorithm.
Clearly, White has winning chances. But is he winning?
24...Qd5
This can hardly be called a good move. A modest rook move like 24...Rf7 was stronger. Then, after the probable (but,
for White, not obligatory) exchange of queens, Black gets counterplay on the d-file, now cleared of pawns – a
dangerous circumstance for us.
25.Qxd5
Just so! The enemy queen has gotten dangerous. From d5 it not only threatens the a2-pawn, it also acts against the white
kingside. In short: White’s king is in danger! From that comes our reply – instant “Petrosian”!
25...cxd5
After the queen trade, it is very important to examine the position once more, with fresh eyes. An algorithm drift is
possible since, after the trade, the sum of the mobility of the players’ pieces could change sharply. In addition, a change
in the safety factor is likely.
In this position, we have m = 1, t = 25/19 = ~1.32 > tcr, approximate parity in the safety factor, ∆k << 0, ∆(25...cxd5) =
~-0.33.
Our diagnosis is a narrow band between the algorithms of Tal and Capablanca. Or the quiet sort of “Tal,” where its
second element is predominant. Or maybe the aggressive sort of “Capablanca.” By playing according to “Capablanca,”
we will place our pieces on the squares that facilitate our attacks on our opponent’s material targets.
Our TC Algorithm differs from the TC Algorithm in the position after 23...Kg7. There – a great “amplitude;” here – not
so much...
What happened in the game?
26.Re6
The best spot for the rook is on either e6 or c6. From there, the rook pressures not only the f6-pawn, but also the
queenside pawns.
26...Rf7 27.Kf1
27...h5 28.h3
Of course, White refuses to be provoked – if 28.gxh5, then 28...Kh6, with powerful counterplay.
28...hxg4 29.hxg4 a5
Or 29...Rd7 30.Ke2 Kf7 31.Kd3 Re7 32.Rc6, declining to trade major pieces. Clearly, exchanges only benefit Black –
compare the density of packing of the white and black pieces.
30.a4
Slightly better was 30.Ke2, followed by Kd3, Rc6, Kd4, f3, and a4. There was no reason to fear 30...a4. Why?
Adopting the plan a2-a3 + b2-b3 + c2xb3 (after exchanging the black a-pawn for the white b-pawn), a3-a4, b3-b4, and
a4-a5, White gets what he was aiming for – the pawn is unstoppable!
30.a4 is “against the rules,” since it is not a piece move. Remember that, in complete accordance with the strategic
Capablanca Algorithm, it is preferable to begin by optimally deploying our pieces and only then advance the pawns.
Point 2 of the strategic algorithm follows Point 1!
32.Kd3 Rd8
Useless! Better to go down fighting than to die on your knees! 32...Re7 33.Rc6 Re4, with counterplay – a double attack
on the pawns at g4 and a4 – is better. And how could we not think of White’s mistake on move 30!
To be fair, though, I should note that all of Black’s counterplay cannot save him. After 34.Rxc7+ Ke8 35.f3 Rxa4 36.c3
(pointed out by Fritz), White still wins, but he must do so in battle!
36...Ke7 37.b4
That’s it!
The parameters of the position are: m = 1, t = 13/8 = ~1.62, approximate parity in the safety factor, ∆k = 0 (!!), ∆(37.b4)
= ~1.12.
Black’s defense falls apart. Why?
Because White has an overwhelming advantage in the second and fifth factors and a long-awaited parity in the density
of packing (compare it with the differences in the ∆k parameter after 23...Kg7 and 25...cxd5).
The conclusion was:
Instead of a postscript to this remarkable game, I offer another new (and, I hope, not too complicated for us) theoretical
concept. We will be speaking of Type I critical positions. In such positions, we have two (or more) moves of the same
practical strength: one in “Tal” style, and the other in “Capablanca” style. In other words, we get the right to choose –
legitimately.
This right to choose is only tenuously connected with our chess brains. Freedom of choice is given to us from above.
That’s how the world is; that’s how the Universe is put together, and the immense complexity of the game is merely the
consequence of the inexhaustibility of the world around us. There are positions which eons of intensive labor, by
thousands of chess geniuses, cannot investigate completely: take merely the starting position of a game!
Type I critical positions – in such positions, we have two (or more) moves of the same practical strength: one
in “Tal” style, and the other in “Capablanca” style. In other words, we get the right to choose – legitimately.
The initial position in our last example, of course, is much simpler than that, and we have some hope of solving it – that
is, to find an answer to the question of which move is stronger, 24.Qd3 or 24.Qe3. It’s a small matter; to find a single
answer we would have to calculate all the variations of the pawn ending after 24.Qd3 Qxd3 25.Re7+ Rf7 26.Rxf7+
Kxf7 27.cxd3. Can we do it during the course of the game? Yes or no?
My advice: don’t torture yourself needlessly in positions that are too complex for you. When selecting a move, trust
your intuition. Or simply roll the dice! Know, dear reader, that – from the viewpoint of contemporary physical
knowledge – in Nature, chaos dominates over order. And that means that the logical dyad “yes” or “no” that we know
so well, needs to be expanded into three. The third element is “indefinite.”
30.Rc2
30...Rab8 31.Rdd2
Karpov is practically pacifistic. He elevates his position in the most modest way possible. And (very importantly for us)
Karpov does this by thinning out his rearguard – that is, by improving the pieces standing at the rear of his fighting
formation.
31...Ne7 32.Kg2
Very consistent.
32...Qa5 33.h4 Rd7 34.Be2
Clearing the fourth rank for future rook maneuvers. In other words, only now does White undertake active operations.
And not for nothing: compare the positions of the centers of gravity of the two sides, and their compactness!
36.Qxd4 Nd5
A tragic oversight.
37.Rxc6
Positions in which the stronger side has an equivalent choice between “elevating” the position by improving
its rear guard or its advance guard, are “Type II critical positions.”
What to do?
Answer: Don’t torture yourself unnecessarily. Believe in intuition, believe in your sympathies and in your feel for
beauty. And don’t be afraid of freedom. Remember: freedom of choice was sanctioned by Caissa. She rolls the dice,
respecting and trusting in His Majesty, Chance.
No. 89: Grischuk – Volokitin
World Team Championship, Beer-Sheva 2005
11.Bxe5
Simple and strong. 11.Nxe5 leads to dangerous and unnecessary complications: 11...Qg5 12.g4 Nf4 13.Nf3 Qh6, with
good compensation for the pawn (as indicated by Rybka).
The “attack” with 12...Qg5 doesn’t work: Black drops the queen after 13.Nf3 Qxg2 14.Rh2.
13.Qf3
13...Qg5 14.g3
Played with the idea of trading the h3-pawn for the c6-pawn?
14...Nxh3
15.Nc4
Amazing! White sidesteps the complications arising from 15.Nxc6 Bd7 16.Qe3. Extend this interesting variation. White
is better here, but why risk it if you don’t have to?
With simple play, White soon obtains an overwhelming position “for free.”
15...b5
“After 15...Be6 White needn’t rush into the attack with16.Qg2 Nf4 17.Qh2 Nh5, as the preliminary 16.Ne3! is far
stronger. And if 15...Qe7 the white pawns are set in motion: 16.Qg2 Ng5 17.f4 Ne6 18.e5! followed by Nc3-e4”
(Shipov).
Six white pawns and their king stand within the rectangle b1-b4-g4-g1, while the black pawns and king are in the
rectangle a8-a4-h4-h8. Thus our ∆k >> 0, as 7/24 >> 7/40. Moreover, White has gained attacking prospects on the
kingside – that open h-file! And finally, White’s hegemony in the center is simply overwhelming.
We may confidently state that Grischuk didn’t give up anything when he refrained on move 15 from the “natural”
capture of the c6-pawn.
18...h5
No salvation is visible.
An esthetically flawless position. It’s almost symmetrical, in regard to the e-file, while the e5-pawn pokes into the
position like a spear. The e5-pawn eyes e8, its queening square.
One more thing: notice the activity White has created for his rooks. They have yet to make a single move (nor will they
be making any), but what a pair of rooks!
Our quickie assessment after 23...Bb7: m = 1, t = 44/36 = ~1.22, “-“, ∆k > 0 (7/30 > 7/40), and ∆(23...Bb7) = ~0.33.
“Tal,” of course!
Now it only remains for White to quietly, accurately strike at Black’s position:
24.g4
24...Nxd4 25.Nxd4 c5 26.Qxb7 Qxd4 27.Qf3 Qxb2 28.Kf2 Rad8 29.gxh5 Rd2+ 30.Kg3 g5 31.Nc4 gxf4+ 32.Kg4
Qxc2 33.Nxd2 Qxd2 34.Qxf4 1-0
13.Qc2
We can do it this way, too! However, were it not for this move, 13.Bxc3 would also be entirely appropriate: after
13...Qxc3 14.e3 0-0 15.0-0 Bxd4 16.Rc1 Qd3 17.Be4, White is better (Fritz).
Note by Rajković: “ ...better was 14...Qc5 15.e3 Bf6 (15...Bxe3? 16.Qxc3! Bxf2+ 17.Ke2 Nc6 18.Qxg7 Rf8 19.Ba3
Nb4 20.Rad1+-; 18...Bd4 19.Bxd4 Qxd4 20.Nd6++-) 16.Bxc3 Bxc3+ 17.Qxc3 0-0 18.0-0².” These annotations were
given in the distant, pre-computer past.
Your task is to extend, and even slightly correct, those notes with the aid of one of our silicon friends.
17...Nd7
The root of all his future terrible suffering, ending in death by suffocation – Zugzwang. The computer advises 17...Nc6
… ...Bd7 or even 17...a5 … ...Ra6.
18.0-0 Rb8
If it were Black’s move in this position, he would play 19...b5 and then 20...Bb7, with complete freedom. But it’s
White’s move!
In this position we have: m = 1, t = 46/28 = ~1.64 (!), “=”, ∆k = 0, and ∆(18...Rb8) = ~0.33.
Diagnosis: “Tal”! Forward!
19.Qa7
An impressive move! By elevating his queen, White not only sticks to the major target on b8, but also completely
paralyzes his opponent’s whole queenside – the second point of the Tal Algorithm in action. Remember that the second
point of this algorithm is at the same time the first point of the Capablanca Algorithm. The line between those two
algorithms is fuzzy...
19...Qc5
After the exchange, a new setup. The best squares for White’s rooks are c1 and d1. These are the best, for now – and
then, don’t worry, Rajković will find a way to “elevate” these pieces.
23...b5 24.Na5
White steadily increases his ∆(move). The evaluation of the rook’s position on d6 is unaffected after 23...b5, since the
rook, while continuing to pressure the enemy knight, now presses on a new target – the a6-pawn. I repeat: the white
rook stands ideally on d6!
24...Nf6 25.h3
White not only elevates the pawn, but also denies the g4 square to the black knight. In other words, White plays to
restrict the mobility of his opponent’s pieces.
25...h6
Because the white pawn’s march to e5 is unstoppable, Black prepares a retreat for the knight.
26.e4 g5
Now the second rook also stands ideally. It’s safe – no fork threatens from e8.
Black starts to choke.
28...Kg7 29.b4
Picture-perfect play! White calmly keeps elevating his position. This is how the great Capablanca himself would teach
us – mortal chessplayers that we are – to play in strategically superior positions.
Why did White elevate his king from g1 specifically to e2? Because the position with the king on e2 is more esthetically
attractive. In Rajković’s play, meta-chessic considerations are clearly uppermost. He knows that he is risking nothing
so long as he keeps improving his position. He is in no rush, openly enjoying his position. Success is guaranteed!
Complex Examples
46.Bf6
46...g6
20.h6
A very modest pawn sacrifice, impossible for most flesh-and-blood players to understand. What’s the idea?
20...gxh6
Loses at once with a deafening crash. True, Anand’s reply was practically impossible to guess over the board.
The cool-headed computer, lacking any sort of emotion or prejudice, prefers two other moves:
a) 20...Nf4 (Rybka’s second line) 21.hxg7 Qxg7 (21...Rg8 22.Bb2 Qc5 23.Qd2+-) 22.Re4 Nxd3+ 23.cxd3+-.
b) 20...Nxe3 (first line) 21.Bxe3 Qe5 22.hxg7 Rg8 23.Qg1 (23.Bh6 Qh2!) 23...Bxf3 24.Kxf3± (according to Anand).
The latter variation can be improved: not 23.Qg1, but 23.Qc1 (Rybka) 23...Bxf3 24.Qa3 and wins, because after
24...Qh2+ 25.Kxf3 the king escapes the black queen’s pursuit by running to c1. There is no perpetual check – nor can
there be one!
The variation 23...Bh3 (… 24...Qh2+ etc.) 24.Qh1, and if 24...Bxg4, then 25.f4!, is utterly entrancing. Don’t forget the
attacking value scale! The queen hangs, and mate in two is threatened. A terrible crush!
But let us return to the game. I am sure that you will be shaken...
21.Bg6
Taking the queen leads quickly to mate: 21...Qxd1 22.Rxe6+, etc. 21...Nxe3 also loses in short order: White continues
with 22.Bxf7+ Kf8 23.Qxd4 Rxd4 24.Bxe3 and Black can resign with a clear conscience.
The strongest move in the diagram position is 21...Qf6 (Rybka), but here too, Black’s situation is not to be envied. One
winning line is 22.Bxf7+ Qxf7 23.Rxf7 Nxe3 24.Qxd8+ Kxd8 25.Bxe3. Extend this variation, please!
21...Ne7
Black is doomed.
23...Rd8 24.Rxd8+ Kxd8 25.Bd3
42.f4
The immediate 42.b5 looks better – White, being in Petrosian territory, is entitled to make a trade first, even if ∆k < 0.
This is because the number of material targets has decreased. The fewer the pieces, the fewer the objects of attack.
And besides (speaking concretely here), following 42.b5 it’s easy to see that Black has nothing better than 42...a5, since
after 42...Kb6 43.bxa6 Kxa6 44.Kb4, and also in the case of 42...axb5+ 43.Kxb5, White – wasting no time – trades his
a-pawn for the enemy d-pawn and then invades the kingside with the king. Victory!
By inducing 42...a5, White frees herself for active operations on the kingside and in the center.
42...Kb6 43.Kd4
And here too, 43.b5 would be totally adequate – after 43...axb5+ 44.axb5 White must win. Check this out, please – no
computer!
43...Kc6 44.Ke3
If Black stands still (45...Kc6 46.f6 Kb6 47.Kd4 Kc6 48.Kc4 Kb6), then White wins with 49.a5+ Kc6 50.h3
(Averbakh), with the normal Zugzwang for pawn endings (see also the postscript to this game). Black is helpless:
50...d5+ (otherwise, after the king retreats to the seventh rank, White plays 51.b5+-) 51.exd5+ exd5+ 52.Kd4 Kd6
53.h4 (Zugzwang again!) 53...Kc6 (53...Ke6 54.b5+-) 54.Ke5 Kb5 55.Kxd5 Kxb4 56.Kd6 Kxa5 57.Ke7, etc.
So instead of standing still, Black needs to play actively: 45...exf5 46.exf5 gxf5 47.Kf4 Kc6 48.Kxf5 Kd5 (… 49...Kc4,
50...d5, etc.). Now 49.b5 axb5 50.axb5 Kc5 51.Ke4 d5+ (51...Kxb5 52.Kd5+-) 52.Ke5 d4 53.b6 (53.Ke4 Kc4±)
53...d3 54.b7 d2 55.b8Q d1Q:
45...a5
50.h3
This is laudable. White has no need to hurry. She wants to push the pawn to h5 when the opposing king is on b6. This
would be good for White, since it brings Black’s counterplay down to nothing. In truth, however, we should note that
h4-h5 also wins with the black king on c5. Here’s a variation: 50.h4 Kc5 51.h5 d5 (otherwise 52.h6, and we have a
position from the game) 52.exd5 gxh5+ (52...Kxd5 53.h6 … Kf4-f7+-; 52...exd5 53.hxg6 hxg6 54.Kf3+-) 53.Kxh5
Kxd5 (53...exd5 54.Kh6 d4 55.Kxh7 d3 56.g6 d2 57.gxf7 d1Q 58.f8Q+ and it’s check!) 54.Kh6 e5 55.Kxh7 e4 56.g6
e3 57.gxf7 e2 58.f8Q and wins.
White’s strategic accomplishments are unquestionable: 6 = 6 (mobility of the pieces; see also the postscript) and
∆(53...Kb6) > ∆(41...Kc7). Indeed, ∆(53...Kb6) = 34/7 – 20/7 = 2.00 > 1.14. What to do next?
Increase the ∆(move)!
54.Kf4
Preparing e4-e5.
54...Kc5
Maybe retreating to the seventh rank was subtler; for example, 54...Kb7, so that after 55.Ke3 Kc7 56.Kd4 she could
forestall the opposing king with 56...Kb6. Now it would be bad to rush with 57.e5, due to 57...d5 with a draw.
However, 57.Kc4 wins, for example 57...Kb7 58.e5 dxe5 (58...d5+ 59.Kc5 transposes to the game) 59.Kd3 Kc7
60.Ke4 Kd6 61.b6 (breaking up the mutual Zugzwang) 61...Kc6 62.Kxe5 Kxb6 63.Kd6 e5 64.Ke7 (64.Kxe5 Kc5 is a
draw!) 64...e4 65.Kxf7 e3 66.Kg7 e2 67.f7 e1Q 68.f8Q with an easy win.
57.e5
The line of least resistance in a strategic attack correlates with the purposeful increase in ∆(move). The problem is the
lingering concern over the protected passed pawn.
57...d5
If 57...dxe5 58.Kxe5 Kc7, the winning line is the forceful 59.b6+, breaking up the mutual Zugzwang. After 59...Kxb6
60.Kd6 e5 61.Ke7, White wins (see also the note to 54...Kc5).
59...Kb7 60.Kd6
64.f7
White’s efforts bear fruit. The potential energy of the far-advanced white pawns guarantees the opponent’s defeat.
After 17...Re8
White has m < 1, t = 50/35 = ~1.43, and an overwhelming advantage in the safety factor.
Diagnosis: the Tal Algorithm.
18.Rxh7
18...Kxh7 19.Qf6
Outstanding! White threatens a primitive mate on f7 after 20.Ng5+, and even Rybka sees no satisfactory defense.
19...Re5
The computer has no sense of humor, but please tell me, can we take its first line seriously: 19...Kg8 20.Ng5 Qe6,
giving up the queen?
It’s a difficult position for Black, very difficult.
20.Kd2
Beautiful! White brings the other rook into the attack, with the threat of 21.Rh1+ followed shortly by mate. Although
the prosaic 20.Qxf7 wins even faster – check it out!
20...Qg4 21.Rh1+
Is that it?
23...b1N+
25...Nb1+ 26.Kb2
24.dxc6
The sacrifice of a piece is clearly a “Tal”-style move. However, the bishop retreat 24.Bd7 looks no worse. In that case
there would probably follow 24...c5, with more or less peaceful, “Capablanca”-style strategic play. In other words, we
have a Type I critical position, where both the bloody and the quiet scenarios are equally probable.
24...bxc6
The bishop is poisoned: 24...gxf5 25.Nd5 Qd8 26.Rb3, and if 26...bxc6, then 27.Qa6 with inevitable mate (Rybka).
25.Bd7
A carefully hidden trap, which even the all-powerful Rybka didn’t see right away.
25...Qc7
Losing. It was necessary to tuck the rook against the lone king – 25...Rb8, and then the heavenly thunder would not have
rumbled...
26.Bxc6+
White’s idea will become understandable perhaps only after four or five moves. For now, we will be formalistic: the
bishop move is the only way to give check here. Let’s remember our eternal attacking mantra: king, queen, rook...
26...Qxc6 27.Nd5
27...Bd8 28.Rc3
28...Qb7
29.Rb3
White insists.
29...Qc6
On 29...Qc8, White wins by 30.Qb5 (threatening mate after 31.Nc7+) 30...a6 31.Qb4 Ka7 32.Rdd3+- (Rybka).
30.Rdd3
Probably the most difficult move to find in this whole brilliant attack – White brings up the final reserve. He threatens
the crushing 31.Rdc3.
30...Ba5
Futile resistance...
31.Rdc3
An innocent trade of a not-very-active rook for the hyperactive black bishop? No answer necessary!
31...Bxc3 32.Qa6
A beauty! I doubt that even our titan, Mikhail Tal, could have done any better.
32.Kg3
32...Rce8
To be or not to be?
Not to be.
17...d5
We have no serious objection to this move, as Black frees himself from all possible and impossible unpleasantness
connected with the pressure by White’s light-squared bishop on f7. The only drawback is the bad bishop on c6.
The moves 17...h6 and 17...Kg8 also look good.
Moving the h-pawn offers instant relief against the enemy attack along that file. However, this could become an
unfortunate tempo loss, since it literally shoves the white dark-squared bishop to the central square d4 (18.Be3 … Qd3,
Bd4). The bishop stands well there, and that’s bad for us!
A rough variation from Rybka: 17...h6 18.Be3 Rde8 19.Bg1 (if 19.Qd3, then 19...Ng4 with good play) 19...Kg8 20.a3
(White plays strategically – the strategic branch of the TC Algorithm). And now if 20...Rc8, then 21.Qd3 … Rfe1, Bd4
is possible. White is better. A position with no counterplay.
There are pros and cons to 17...Kg8, too. Here it is good that Black immediately – and, what is important to us,
significantly – increases the local density of the pieces around his king. The drawback is that White gains the possibility
of elevating his rook to the third rank, thereby obtaining real chances for an attack on Black’s king.
However, 17...Kg8 can hardly be of independent significance. With the king on g8, Black probably cannot do without
...d6-d5, meaning that he has every prospect of reaching the game position (see 17...d5 and 20...Kg8).
18.Qh4
White intends to put a rook on h3, trading off the dark-squared bishop and mating on h7. Good intentions – destined to
fail!
18...Rfe8
Andersson plans to sacrifice the h7-pawn. The goal is to create a safe haven for the king over on the queenside. A
tremendous idea!
19.Rd3 Qc5
Drawing closer to the king. And not just to the black one, but also the white one – Andersson quietly begins to weave
intrigues against the white monarch, who appears to be in complete safety.
With the help of a true and emotionless silicon friend, I hasten to note that after 19.Rd3, Black has a one-hundred-
percent chance to put a lid on the opponent’s kingside offensive with 19...d4!. Now the attack up the h-file is not
possible: 20.Rh3 dxc3 21.Bxf6 Qxf5, and Black wins! Effective, no?
White would have to take this pawn with 20.Rxd4 and shift from attack over to defense – a complete turnabout!
Unfortunately, analysis shows that Black doesn’t get full compensation for the sacrificed pawn.
20.Rh3 Kg8
Wrong! White wins a pawn but loses the initiative. He should have played according not to “Tal,” but to “Capablanca,”
trying to improve his position as much as possible. Why?
Because the CP Algorithm for Black automatically indicates the TC Algorithm for White. And the TC Algorithm is
“Tal” and “Capablanca” at the same time.
The “paradoxical” move 21.a3 (Rybka’s recommendation) deserved serious attention. White intends to retreat the
bishop to a2 and follow up with b2-b4, Ba2-b3, a2-a4, and b4-b5, increasing the ∆(move) of his position. And if 21...h5
(again, recommended by Rybka; to me, it seems almost pathological), then shifting the knight to f4 with Nc3-e2-f4
looks very strong. Black has serious problems.
A different line of defense after 21.a3 is ...a6-a5 and ...b7-b6. Black now elevates, not the h-pawn, but the pawns on the
queen’s wing. But then, let’s say after 21.a3 a5, White pushes the pawn to a4 (that weak point b5!) and starts a slow,
strategic game with the h-pawn’s execution reserved for later.
After 21.a3 White is no doubt better.
23.Qh8+
This move doesn’t seem to spoil anything yet. Still, 23.Rd3, sounding the retreat without delay, is for choice. The d3-
rook is clearly better used in defense.
The following long – and therefore not completely indisputable – variation is possible: 23.Rd3 Re5 (Stohl’s engine-
endorsed recommendation) 24.Qh8+ Ke7 25.Qh5 Rd7 26.Ne2 (26.Na4 Qa5 27.Nc3 Qc5 28.Na4 – draw?) 26...Bb5
27.Nf4 Qc6 28.Rxd5 Rdxd5 29.Nxd5+ Rxd5 30.Bxd5 Qxd5 31.Re1+ Be5 32.Qg5+. And everywhere = (0.00) – at 13,
14 ply (half-moves)... A position of dynamic equilibrium?
Answer this question on your own – of course, with the help of the computer.
A step towards the precipice. Against the needs of the position, White plays for the attack, and that leads to catastrophe.
Preferable is 25.Rd3 (… Rdd1), giving up a pawn, for example after 25...Kd8 26.Rdd1 Bxc3 27.bxc3 Qxc3. On the other
hand, Black is not required to rush into trading his bishop for the knight. He’s better (³), but nothing more than that.
The strongest move (Rybka-approved) is probably 25.Na4. An illustrative variation is 25...Bxa4 26.Re1+ Kd8 27.Rxe8+
Kxe8 28.Bxa4 b5 29.Bb3 Re7 30.Qd1 Re4 31.Rf3 Bxb2 32.Rf1, with dynamic equilibrium – White has neutralized his
opponent’s initiative. One important consideration is that, after 28.Bxa4, a potential assassin of the white king (the
light-squared bishop) leaves the board – whereas in the actual game it stays alive.
25...Kd8 26.Rxe8+
With this trade, White catastrophically weakens his first rank. Mistake after mistake!
26...Kxe8 27.Qh8+
27...Ke7 28.Nd1
And here retreating the queen is imperative – 28.Qh5, with at least a little bit of hope...
28...d4
29.Qg8
The apex of White’s suicidal idea – the target on f7 is attacked. Clearly an unwarranted act.
29...Qe5
A just and harsh punishment for White’s previous missteps. Mate is threatened, and there’s no real defense to this
simple threat.
33.Qe1 Re7
White resigned without waiting for mate in two (two different ways).
Andersson’s play leaves a very strong impression.
20.f4
Mikenas: “I saw this possibility well before this, but for a long time could not decide on it. The point was that such a
sharp continuation requires exact calculation. Carefully examining all the possible variations, still I embarked upon this
step.” (I am citing from the book Vladas Mikenas, Fizkul’tura i Sport, Moscow 1987.)
As will become clear a little later, with this fearless move White sacrifices not just a pawn, but also a rook. Get ready to
see fireworks!
Computer analysis shows that the sacrifices are correct, and should have led (with correct defense) to a considerable
plus for White. However, the same computer analysis (thanks to Rybka) calmly also indicates a more powerful move:
20.Qd3, when after 21.Ng4 White intends to force the win immediately.
The variations are:
a) 20...Kg7 21.f4 Qd8 22.fxg5 hxg5 23.Qe2 Qe7 24.Rxf6 Qxf6 25.Ng4 Qe7 26.Be5++-.
b) 20...Qd8 21.Qc2 Kg7 22.Rcd1 Qe8 23.Ng4 Nxg4 24.Qh7+ Kf6 25.f3+-.
I’d like you to extend these variations, and then either to condemn or approve of them.
20.f4
20...Bxe3+
21.Kh1 Bxc1
Or 21...Qe7 22.Qc2 Rd8 23.fxg5 Bxg5 24.Bf4, and White must win (this variation is from Rybka).
22.fxg5
Too beautiful, too elegant! Evidently Mikenas, entranced by his idea, was full of sacrificial fervor: he wanted to get at
the target on f6 by the shortest route possible (the e5-knight is more valuable to him than both rooks!), and then mate
on h7!
To carry out all this would not be easy. I add that there is a hole in Mikenas’s analysis – see his vast notes to Black’s
move 23.
The simple 22.Qxc1 (as offered by Rybka) was easier: the computer sees no salvation. Here’s one line out of many:
22...Nh5 23.fxg5 Qb5 24.Ng4 Nxg3+ 25.hxg3 hxg5 26.c4 Qc5 27.Rf5 and White wins. This variation is fairly long
and so requires additional research. Dare to plunge into it! And let your project be aided by ultramodern chess engines
– chess is truly inexhaustible.
22...Bxg5
23.Rxf6
23...Kg7
But with this one, I have a thousand. The king move – if we are to believe Rybka – leads to forced mate in sixteen
moves against best defense!
A word from Mikenas. Here is what he writes in his notes to the diagram position (remember, these were pre-computer
age commentaries): “The rook cannot be taken: 23...Bxf6 24.Qd3, and mate is unavoidable. But Black could have put
up a more stubborn defense with 23...Qb5! 24.Qc2 Qxb1+!! 25.Qxb1 Bxf6, getting two rooks for his queen. True, this
variation is not forced. I intended to continue with my kingside attack, although I was not convinced of the correctness
of my calculations. It was only from mutual analysis after the game that we found an interesting variation of the attack:
24.Ng4! Bxf6 (there is nothing better: White threatens 25.Qc2) 25.Nxf6+ Kg7 26.Bh4! Qc4 27.Be4 e5 28.h3! (there’s
no sense in White hurrying into the immediate 28.Qe1, since after 28...Bg4 29.Nxg4 f5 Black defends successfully)
28...Bf5 29.Nh5+ Kg8 30.Qf3! Bxe4 31.Qf6 Bxg2+ 32.Kh2, and White wins.” I add: if only it were so.
A bucket of cold water: after 30.Qf3! (Mikenas’s mark) there is 30...Bg6 and Black is better! More likely, ³ and not-+
(which one is irrelevant to us). Alas! Black glues the bishop and the king together, and soaks up the pressure.
However, we too have good news from Rybka: after 27...e5 White must continue, not 28.h3, but 28.Nh5+-. The easy
solution lies right on the surface! Here’s a simple variation: 28...Kh8 (28...Kg8 29.Qf3+-) 29.Bf6+ Kg8 30.Qf3, and
Black is helpless.
That’s it: the attack was correct after all. Fortune favors the bold!
The game’s conclusion leaves an indelible impression.
24.Qd3 h5 25.h4
36.Rdg1
A natural move – but maybe not the strongest one. Rybka prefers 36.Qe2, allowing White (after, let’s say, 36...Qb2) to
take the queen to h5. Then White brings his rook to h3 with extremely uncomfortable, permanent pressure on Black’s
kingside. Opposite-colored bishops in the attack – that’s always a plus, and often a very serious one!
36...Qb2
Defending not only against 37.Qe2 … Qh5, Rh3, but also against 37.Qd4.
37.Rg5
After this flexible, White gets not just to mass all his major pieces on the g-file (with 38.Qg3), but also to transfer the
queen to h3 (38.Qh3 … 39.Rh5).
37...Qf6
38.R1g4
38...Qa1+ 39.Kg2
Let’s examine the position after 39.Kg2 from Black’s point of view. What can he do – how does he play here? Or,
finally, what does he have?
Answer: m > 1, t = 43/38 = ~1.13 and an indefinite safety factor. I add that Karpov has improved his pieces – that is, he
has made significant additions in the fourth and fifth parameters, although not for free. The price is a deficit in the “t”
parameter and also in the third parameter. It looks like the World Champion, by that very fact, has lost part of his
previous advantage.
Our diagnosis of this position is: the TCP Algorithm. That means that Black must start with “Tal.”
39...Qb2+
39...Qxa5 loses to 40.Qd4. Now Black is saved neither by 40...Ree7 41.Rxg7 Rxg7 42.Qf6 Qd2+ 43.Kh3 Qe3+ 44.Rg3,
nor by 40...Qa2+ 41.Kg3 Qb3+ 42.Kh4 Qa2 43.h3:
Position after 43.h3 (analysis)
For us, it is both important and interesting that ∆(43.h3) = ~1.32 in the diagram position. To this huge ∆(move), add
White’s overwhelming advantage in the third and fourth parameters.
But after 39...Qb2+ Black has every prospect of holding. How?
40.Kh3 Rce7
41.f5 Qf6
But this move is confusing, to put it mildly. Better is 41...Qa1, threatening to check from the rear. Black needs to play
actively! Why?
Because (sorry for repeating this) the TCP Algorithm starts off with “Tal” and not “Petrosian.”
Karpov himself indicates 41...Qa1, and he also points out 41...Rf8 (the first and second lines by Rybka).
42.Rh5
Of course! Now Black’s queen is hobbled – 42...Qa1 43.Rxh7++-. White intends to place his other rook on h4,
provoking Black into moving the h-pawn.
42...Rf8
We probably do not have the right to criticize this move – Black has accepted that ...h7-h6 is inevitable, and intends, in
the variations, to push the rook to f6 in order to securely guard the target on h6.
The attempt to keep the pawn on h7 most likely will fail. In support of this we offer a long variation given by Rybka:
42...Rc7 43.Rgh4 Bf8 44.Qg3 Bh6 (bravo!) 45.Rxe4 (45.Rxh6 Qxf5+ 46.Rg4 Rg7 47.Qf4 Qxg4+ 48.Qxg4 Rxg4
49.Kxg4 e3 50.Bf3 e2=) 45...Rcc8 46.Reh4 Re3 47.Qxe3 Qxh4+ 48.Rxh4 Bxe3 49.Rg4 h5 50.Rg6 Kh7 51.Rxd6, and
White has every chance to win this.
Of course, one must take long variations with a grain of salt. Try to improve both sides’ play.
A practical player cannot calculate such a long variation at the board. He trudges along a thousand times more slowly
than a modern chess computer. He moves by feel, often making decisions out of general considerations. And these
general considerations could be maxims of the strategic algorithm.
When playing strategically, we must hew closely to the third, fourth, and fifth parameters of the chess position. Our
idols are safety, compactness, and spatial expansion!
43.Rgh4 h6 44.Rg4
In the diagram position, Black has a lot of possibilities – there is no threat of a triple capture on h6, the queen is free (or
almost free!) to maneuver. In short, a breather – but... But his position is seriously compromised – there is now a hole
on g6!
I won’t irritate the reader with an endless string of variations prepared by our hardware helpers. I will think of schemes
(setups), for instance Rg6 + Kg4 (we have already talked about this) and... Kh1 + Rg1 (saving the king from the
irritating check on a1) +, let’s say, Bd5-e6, Qe3xe4, etc. I do not intend long calculations – really! I have faith, I will go
ahead by feel...
44...Re5
45.Rgg5
One more baby step toward the goal – White increases his ∆(move).
45...Rc8
A “Tal”-style move, as Black threatens to win after 46...Rxd5. Black can’t play this way, though. He has to to play
according to “Petrosian” and not “Tal”!
46.Kg4
46...Kh7
47.Rg6 Qf8
Do you believe in the merciless power of the strategic principle we have endured, of increasing the ∆(move)? If, for
now, you do not believe in it, I hope that White’s next simple move will make you believe ...
Karpov replied
48.Qg5
Marvelous! The threat is 49.Rgxh6+ Bxh6 50.Qg6+, with a quick mate. Black is helpless.
Nunn resigned after
The shortest afterword we can think of to this amazing game is ∆(48.Qg5) = ~1.96.
22...h6
In the most brazen way possible, Black has broken the one unshakable ancient rule that says the weaker side should
never move pawns where his opponent is attacking.
A mistake? No, no, and again no! Svidler has envisioned the transformation of one defensible position into another. He
(see below) could maintain approximate parity in the third factor.
But in the language of “Capablanca” (we won’t overstate the danger to the king’s position), Svidler’s move could even,
on the whole, be considered trivial: Black, playing strategically, is totally within his rights to make a trade first, given
that ∆k = 0. And so Black fearlessly “elevates” his position. Bravo!
On the natural 22...Rac8 (instead of the “counterintuitive” 22...h6), there would most likely follow 23.Rd2 Rc7 24.Rhd1
Qc6 25.Bg2 Qa4 (25...e5) 26.Qe3 Rc5 27.Nc1 Re8 (Rybka). Black stands slightly worse – the white g5- and h5-pawns
hang unpleasantly over the king’s head. A complex struggle...
23.gxh6 gxh6
After the pawn trade, it is very easy to see that White has no serious threats, since the black king will shortly find safe
haven without great difficulty. That haven will be on the h-file. And the square h7 is clearly more appealing than h8.
Why? Thanks to the compactness factor!
The first line from Rybka is 24.Qd3. The second line is 24.f5, as Adams chose; we will examine it later.
After 24.Qd3 Black has two nice moves: the very risky 24...Rad8 and the very solid 24...Kh8 (h7 is “forbidden” – see
that queen on d3!).
After the “very risky” 24...Rad8, I dare to “frighten” you with a not very short (and therefore highly suspect) variation,
and ask you to support it or refute it yourself. Here is this variation, which admittedly I do like a lot: 25.Bxe6 fxe6
26.Nd4 Bc8 27.e5 Rf7 28.Rhg1+ Kh8 29.Qg6 Rdf8 (29...Rh7 30.Nb3 … Rg4, Rdg1+-) 30.Nxe6 Bxe6 31.Qxh6+ Rh7
32.Qxe6 Qc8 (32...dxe5 33.Rg6 … 34.Rdg1±) 33.Qxd5 dxe5, with an approximately equal game.
Of course White, if he wished, could safely reject the sharp 25.Bxe6. He has other candidate moves: 25.Nd4, 25.Rhg1+.
In reply – bishop to f6, king to h8... A complex and, most likely, strategic struggle, as the black king feels quite
comfortable. The evaluation of that position would be ²: compare that with the 22...Rac8 variation.
Our next step is the position after the “solid” 24...Kh8. In order not to drown you, dear reader, in a sea of variations, I
present only a very small number of them. The bare minimum! Beauty will be the guideline for their selection. That
will help us mortal chessplayers to approach the primeval depths of our game.
Quite possibly, on 24...Kh8 Adams would have played 25.Nd4, when Black would have replied 25...Bf6. This is not
just prophylaxis (the bishop comes closer to its king), but also a move bearing the signs of the Tal Algorithm, since
from f6 the bishop hits b2!
“Tal”? Don’t hurry with the reply!
Going deeper: 26.f5 a5 (who’s attacking whom?) 27.fxe6 fxe6:
Position after 27...fxe6 (analysis)
The promised minimal variations:
a) 28.Nxe6 Qe5 29.Qb3 Rfe8 30.Nc7 (30.Nd4; 30.Rhf1) 30...Bc6 31.Nxe8 Ba4. Who is worse?
b) 28.Bxe6 Rae8 29.Bd5 Bxd5 30.exd5 Bxd4 31.Qxd4+ Qxd4 32.Rxd4 Rf5 … 33...Ree5, with good compensation for
the pawn and an almost guaranteed draw.
c) 28.Rhe1 (likely the most dangerous move for Black) 28...Bxd4 29.Qxd4+ Qxd4 30.Rxd4 d5 31.exd5 (31.Bxe6 dxe4
32.Rd7 Rf6²) 31...Bxd5 32.a3 bxa3 33.Bxe6 Bf3 34.Bb3², the same evaluation as before.
The general conclusion: 24.Qd3 (the first line by Rybka) doesn’t refute Svidler’s idea.
Almost certainly, White also doesn’t have anything serious after the game move.
24.f5
24...Qe5
A multipurpose move! Black not only brings the queen over to his half-exposed king, but also attacks enemy targets on
e4 and, most importantly, on b2. We mustn’t forget that on the right side (a1-a8-c8-c1) Black’s forces have and will (at
least in the near future) continue to enjoy a great superiority. Black has the right to counterattack!
Many hours of computer analysis showed that Black holds this position securely. I add that it’s a very difficult position
for an analyst, who is looking for the one strongest move. In such a sea of variations, even Rybka feels uncomfortable
and changes its evaluations frequently.
As difficult as this is for a computer, it was still harder for Adams and Svidler – two gifted chessplayers, world
championship candidates. All the more so for us mere mortals... What are we to do?
Answer: find the strongest move by blindly forging ahead. In this it is sometimes necessary to proceed from general
chess considerations. And you, dear reader, must always work with convincing (though short) variations. Let your
concrete tactics be at the service of flawless strategy!
One more thing – a quote from Capablanca (I quote this, not for the first time, nor for the last): “Never refrain from a
move out of the fear of losing. If you think a move good, better to make it, without thinking of the result. Experience is
the best teacher. Remember: if you want to become a good chessplayer, you must lose hundreds of games first.”
I will continue the theme of searching for the strongest move later, after concluding the analysis to the present game.
25.Rhg1+
Or 25.Rd4 a5 26.fxe6 fxe6 27.Qg4+ Kh8 28.Qxe6 Rf6 29.Qxe5 dxe5, with an outstanding game for the sacrificed pawn
(Rybka’s variation).
25...Kh7
25...Bg5 wasn’t bad, either – the engine’s recommendation. But I like the human’s move better. Why?
Because it assures Black of a greater density of packing of his king and pawns’ position!
26.fxe6
Rybka considers 26.Bg2 strongest here, with 26...Rac8 27.fxe6 fxe6 28.Qb3 Rf2 to follow. But Black has another path!
It would be interesting to work out the more compact 26...Rg8 – does this purely “Petrosian”-like move work? Rybka
replies 27.fxe6 fxe6 28.Qb3 d5 29.exd5 (a “Tal”-style piece sacrifice!) 29...Qxe2 (Black’s king is densely packed, and
therefore he has every prospect of warding off the attack!) 30.Bf3 Qe5 31.Rge1 Qf4 32.Re4 Qf8 (closer to the king)
33.Rxe6 Kh8, and Black is out of danger! Try and refute this, if you can...
26...Bxe4
A quote from Capablanca: “Never refrain from a move out of the fear of losing. If you think a move good,
better to make it, without thinking of the result. Experience is the best teacher. Remember: if you want to
become a good chessplayer, you must lose hundreds of games first.”
The universal method for searching for the strongest move may be filled out by a sort of extra-chessic (supra-chessic,
meta-chessic) component. For us, that will be the esthetic element in the game of chess. We will activate this aspect
when we have a choice between the beautiful move and the not-so-beautiful one, an ugly one.
Why?
Because we believe (and strongly!) that the beautiful move in chess has greater chances of being the strongest one ...
We know a lot of “formulas” for beauty. Here is one of the favorites: simplicity + symmetry + paradox. This formula, it
seems to me, coincides with the formula for the beautiful in our modern love of nature, and we cover it with the
magical veil of our beloved game...
12.Nc6
Only like this, although no doubt the game move required not just courage, but also accurate calculation. By invading c6
with the knight, White risks dropping it. But he is not afraid of this – on the contrary, he intends to sacrifice it shortly!
12...Qd7
“It turns out that 12...Qxd1 13.Rxd1 Bb7 does not work, in view of 14.Nb5! Bxc6 15.Nc7+ Kf8 16.Nxa8, and the
knight cannot be taken in view of the mate on d8” (Kholmov).
13.Nxe7
13...Kxe7
Kholmov: “After this, Black loses quickly. The question arises: was it possible, on the whole, to defend?
“In the first place, a preliminary queen trade suggests itself. And so, 13...Qxd1 14.Rxd1 Kxe7 15.Bg5+! Ke6 (15...f6?
16.exf6+ Bxf6 17.Nd5+) 16.Rd6+ Kf5 (16...Kxe5 is worse: 17.Rd5+! Ke6 18.Re1+) 17.f4!.
“Now White threatens to capture on h6 and give mate on f6, while on 17...Ng8 I had prepared 18.g4#. After 17...Be6,
Black also loses quickly: 18.Ne2! Ke4 19.Ng3+ Ke3 20.Re1#. The only way to hold out for longer was 17...Bxe5, but
then too, after 18.Rd5 f6 19.Bxh6 Bb7 20.fxe5 Bxd5 21.Nxd5 Kxe5 22.c4, White must win, since the f6-pawn is lost,
too. In this variation, Black can try not giving up this pawn by playing 21...fxe5. However, here he would jump out of
the frying pan into the fire – a rather pretty finish follows: 22.g4+! Ke4 23.Nc3+ Kf3 24.Re1 Kg3 25.Re3+ Kh4
26.Ne4:
Position after 26.Ne4 (analysis)
and mate next move.”
More variations (I’m quoting again here): “And if 23...Kd4 [instead of 23...Kf3 in the previous line – A.Sh.] then
24.Re1! c4 25.Be3# is no less effective. Finally, if 19...g5 (instead of 19...Bb7), 20.fxe5 Kg6 21.Rf1 Kxh6 22.Rxf6+
Kg7 23.Ne4, Black’s position is also hopeless.”
These pre-computer analyses by GM Kholmov (see Ratmir Kholmov, Moscow: Fizkul’tura i Sport, 1982) were all
checked by Rybka. No comment needed!
Kholmov considers Black’s best defense to be 13...Qxe7, but here too, as Rybka shows, Black is doing poorly. Some
variations:
14.Nd5 Qd8 15.Nf6+ and now:
a) 15...Ke7 16.Qf3 Be6 17.Bg5 Qc8 18.Nh5++-;
b) 15...Bxf6 16.exf6+ Be6 17.Bxh6 Qxf6 18.Qd6 Rd8 19.Qg3 Rd5 20.c4 Rd7 21.Bg5 Qd4 22.Qb8++-.
And so, 13...Qxe7 loses. Keres’s move, 13...Kxe7, also loses – in every variation!
How, specifically? Let the victor demonstrate.
14.Bxh6 Bxh6
15.Qf3 Bg7
“Black parries the chief threat (16.Qf6+), ignoring the secondary threat (16.Qxa8). It seems that 15...Re8 was stronger,
preparing to retreat the king to f8, but here too White would have a great choice of continuations. I could play 16.Rad1
Qb7 17.Qf6+ Kf8 18.Qh8+ Ke7 19.Qxh7, with enough of an attack to win, but better would be the more energetic
16.e6! fxe6 17.Rad1 Qb7 18.Nd5+ Kd6 19.Nb4+! Kc7 20.Qg3+ e5 21.Rxe5!, with an easy win” (Kholmov).
I add that 15...Bg5 loses, too. White prevails with the “standard” 16.e6 fxe6 17.Rad1, indicated by the computer, and
then the no less “standard” 17...Qb7:
Position after 17...Qb7 (analysis)
18.Rxe6+! wins. Extend this, please – now without a computer!
16.Nd5+
“Of course, 16.Qxa8? would not be good because of 16...Bb7 17.Qxa7 Qc6 18.f3 Ra8 19.Nd5+ Qxd5 with an uncertain
outcome” (Kholmov).
16...Kd8
17.Rad1 Bb7
Kholmov: “Tempting was the attempt to remove his queen from the dangerous file by 17...Qb7, expecting 18.Nxb6+
Kc7 19.Nxa8+ Kb8, with a safe visit by the king, but that runs into the energetic reply 18.e6!!” [indeed, this
“standard” e-pawn move holds the key to the enemy’s fortress – A.Sh.].
Out of the whirlwind of pretty variations offered by Kholmov and supplemented by Rybka, I show possibly the most
beautiful of all: 18...fxe6 19.Nb4 Bd4 20.Rxd4+ cxd4 21.Qf6+ Kc7 22.Qe5+ Kd7 23.Qxe6+ Kd8 24.Qf6+ Kc7
25.Qe5+ Kd8 26.Nc6+:
Position after 26.Nc6 (analysis)
Mate is unavoidable.
How did the game end? Rather boring prose! I note only that Kholmov played flawlessly as the game progressed.
The queen and rook are optimally placed. Isn’t it time to make a pawn move?
24.a4 g5 25.Qd5 Rhe8 26.Rxh7 g4 27.a5 gxh3 28.axb6+ Kxb6 29.Rxd7 1-0
36.Rxg7+
A “Tal”-style move; the die is cast. One alternative would be 36.Rxd3. By giving up the exchange (or giving up his rook
for bishop and pawn, anyway!), White also wins in all variations. For example, 36...exd3 37.f6, and Black is
defenseless – the threat being 38.Qh6 g6 39.f6 with an easy win.
However, there’s another, much more interesting variation – 36...Rb2+ (instead of the primitive 36...exd3) 37.Ke3 exd3
38.f6 d2 39.e7 d1N+ 40.Rxd1 Qxc3+ 41.Ke4:
Position after 41.Ke4 (analysis)
After 41...Qc8 (probably strongest) 42.exf8Q+ Qxf8, the computer insists on 43.Nf5. Guaranteed victory:
a) 43...gxf6 44.Qg4+ Kh8 45.Rd8+-;
b) 43...Qxf6 44.Nh6++-;
c) 43...Qa8+ 44.Rd5+-;
d) 43...Re2+ 44.Kd3+-.
The game continued:
Here Rybka recommends 39...Bc4, but in any case – luckily for us! – adjudicates victory for White: 40.exf8Q+ Qxf8
41.f6 Bf7 42.Ke3, etc. The cold-blooded 40.f6 is also very strong, when after 40...Qxg5+ 41.hxg5 Re8 (nor does
41...Rb8 save Black), we get the same position as in the game after Black’s move 41.
After move 39 in the actual game, the following position arose:
40.f6
Brilliant! “Exchanging queens, a whole rook down, is the last thing that could enter a chessplayer’s head!” (Yudasin)
The fascinating 40.Qf6+ Kg8 41.Qe6+ Kh8 42.f6 also wins, as Black has no defense against 43.f7. At least, neither
Fritz nor Rybka can see one. One rough variation goes 42...Qxa3 43.f7 Qb2+ 44.Ke3 Qc1+ 45.Kf2 Qf4+ 46.Kg2
Qd2+ 47.Kg3 Qe3+ 48.Kg4 Be2+ 49.Kf5 Qf2+ 50.Ke5 Qg3+ 51.Kd5:
Impressive, eh?
We return to the game after 40.f6. There follows the queen trade:
40...Qxg5+ 41.hxg5
And then Black’s attempt to blockade the threatening white pawns:
41...Bc4
What to do? The stunning reply: to play strategically – that is, to increase his ∆(move)!
42.Nf5
44...Rc8 45.c4
45...a4 46.c5
Eyeing c8.
46...R8b8 47.c6
47...Rc8
48.f7
A mistake. That’s at least what Hübner thinks, and the German grandmaster’s opinion coincides with Rybka’s
assessment.
Hübner gives 22...Bf4 23.Rxf6 Kxf6 24.Qf3 Qc7 (this modest queen move is what Karpov overlooked in his
commentary) 25.Rh4 Kxe7 26.Rxf4 Kf8, and Black stands no worse!
And how is it with White? He has complete compensation for his minimal material deficit.
After 26...Kf8 White would more than likely have continued 27.a3 – prophylaxis against the opponent’s attack along
the first rank. And then Ka2, Nf5 hitting the target on h6, with a complex struggle.
The game continued
23.Qe5
Returning the favor! One gets the impression that the World Champion was in a fighting mood and didn’t see the queen
trade. He perceived only one target – the enemy king – and therefore disdained the other, less bloody possibility.
After 23.Qxg4 Nxg4 24.Nf5+ Kh7 25.Re1 Nxf2 26.Rd7 … 27.Nd6 (Rybka), White has real winning chances. Getting
somewhat ahead of ourselves, I add: after the text, White is not better!
23...Kg8
A good move. However, computer analysis shows that the other retreat, 23...Kh7 … 24...Ng8µ, was just as good. So,
after 23...Kh7, there was only one thing for White to do – take the knight: 24.Rxf6 Bxf6 25.Qxf6. But then, to White’s
chagrin, Black would have the very unpleasant 25...Qe6. Karpov evaluates the position after 26.Qxe6 fxe6 with one
symbol: µ. Question: what to do?
Answer: White has the saving check 26.Qf5+, and Black must refrain from trading queens! After 26...Kg8 27.Qxc5 a
hard fight lies ahead. Approximate equality.
Conclusion: 23...Kh7 was enough for equality. Black holds the position.
He also holds after the game move, 23...Kg8. The question is: how?
You will see the answer in my notes to Black’s move 24, but first...
24.Re1
Karpov: “Of course, not 24.Rxf6 Bxf6 25.Qxf6 – 25...Qe6.” Clearly and accurately put – see also my notes to 23...Kg8.
24...Nd7
“In the event of 24...Nxh5 25.Re4 Nf4 26.Nf5 f6 27.Qxf6 Bxf6 (27...Qd1+ 28.Rxd1 Bxf6 29.Rxf4) 28.Nxh6+ Kg7
29.Nxg4 Bxe7 30.Rd7 Ng6 31.f4, White has the advantage, but still stronger is 25.Nf5 (instead of 25.Re4) 25...Bf4
26.Qd5 Ng7 27.Rf6 Qh5 28.g4” (Karpov). An important addendum: that book came out way back in 1984...
This pre-computer analysis we can and must correct. This is what our loyal software assistant discovered:
1) 25...Kh8 instead of 25...Kg7 in the variation with 25.Re4, or
2) 25...Qf4 in the variation with 25.Nf5.
And now, the variations themselves (after my editing):
1) 24...Nxh5 25.Re4 Nf4 26.Nf5 f6 27.Qxf6 Bxf6 28.Nxh6+ Kh8 29.Nxg4 Bxe7 30.Rd7 Bg5=;
2) 24...Nxh5 25.Nf5 Qf4 26.Qd5 Nf6 27.Qf3 Qxf3 28.gxf3 Nh5 29.Rg1 Ng7 30.Nxg7 Kxg7 31.f4 Rxe7 32.fxg5 h5=.
I ask you to correct (or even to refute!) these “final” variations. If, of course, you can do it. A difficult battle, even for
world champions and their challengers...
A tentative conclusion: Black holds the position also after 23...Kg8 24.Re1 Nxh5.
But after the move in the game (24...Nd7), he loses by force.
25.Rxd7
There is nothing supernatural about this sacrifice, since in an attack on the king, chess time (here, a single tempo) is
more important than material (it’s only an exchange).
But here’s a genuine miracle: Karpov voluntarily enters into an ending while down a rook!
Open attack.
29...Bc7
And suddenly...
30.Kc2
30...b5 31.Nxh6+
White doesn’t rush. The d-pawn’s advance, which will decide the game, can wait a bit.
At long last!
Obviously, Karpov doesn’t want to get up from this game. Just like us, his admiring fans! And nonetheless...
17...Qxb2
18.Nd5
This is not an exchange. Though temporary, it’s still a piece sacrifice, because after...
18...Nxd5
19...Qf6
Closer to the king! Note that the knight on d5 can’t move: 19...Nxf4 20.Nxe7++-. That’s no longer the case after
Black’s move.
Curiously, Rybka, that emotionless and esthetically challenged computer program, at first clearly underestimates the
depth of Anand’s conception. It prefers – and I am beside myself with indignation! – 19...g6 and 19...Kh8.
Unbelievable!
Anger aside, I love Rybka. But I (and you, I hope) will not overestimate its powers.
What does the computer lack? It lacks a feel for beauty!
Beauty is a powerful force that helps us flesh-and-blood players to touch the chess Absolute...
20.exd5
The point. Having rid himself of his excess ∆(move) with 19...Qf6 and 20...Bf8, Anand has managed almost completely
to ensure the safety of his king. And when he rids himself of that “extra” pawn on f7, then he will, all in all, be very,
very happy.
21.Nh6+
Or 21.Be3 Qg6 22.Qf3 f6 (Rybka), and Black is out of danger. With an extra pawn!
24.Qh3
Polgár’s attempt to revive her sputtering attack (24.Qh3 … Rf3, Raf1) soon ends in an instructive knockout. The
computer advises 24.Qxg6, going into an inferior, maybe losing endgame.
24...Ne4
Closer to the king – and to the white king too! Along the way, Black threatens to capture the c2-pawn – not with the
queen anymore, but with the c8-rook! Who’s attacking whom?
25.Rf3
I have already told you of White’s “fearsome” plans: 26.Raf1 and... what then?
On the other hand, it’s no crime to honor this heroic attempt, and...
25...h5
...and offer one’s surprised opponent a queen trade. Anand is playing across the whole board! Amazing!
26.Raf1
Clearly played out of inertia, and most likely in severe time pressure. GM Polgár, obviously, has simply overlooked the
simple “mate” to the white queen, after which any sort of resistance is pointless. The “paradoxical” 26.g4 (Rybka)
merits attention, hoping for 26...hxg4 27.Qh8#, or 26...Qxg4 27.Nh6++-!
On the other hand, after the cool-headed 26...Nf6, White loses in all variations. Check it out!
26...Qg4
29...Nd2 30.Ng5 Nxf3 31.gxf3 Rxc2 32.Ne6 Rec8 33.Be3 Re2 34.Bf4 Ra2 35.Rg1 Rxa4 36.Bh6 Kh8 37.Nxg7 Kh7
0-1
Let’s hear from Carlsen (64, No. 12/2009): “One can hardly call 16...b5 17.a3 a solution to Black’s problems, since
Black cannot play ...Bb7 because of Bg4 Qf6 and Bd4, and the queen is trapped.”
I say: 16...Qd7 was worth a look.
17.Rc1 Bd7
18.Bf3
18...Rac8 19.Qb3 b5
If he can’t do without this move (a tremendous weakening of the queenside), then it looks like Black is in a bad way. I
repeat: 15...Qd7, 16...Qd7, and 17...Qd7 all deserved a look...
“The key move! Black’s pieces look stupid on the kingside, and it’s as if the sheaves of pawns on the queenside are
awaiting their reapers” (Carlsen).
Well put!
Looking at the queen standing on that useless square h5, for the last time in this game I will say something extremely
unpleasant for Black, which is that he should have played 15...Qd7, or 16...Qd7, or 17...Qd7...
23...Ng8
A tactical oversight in a very difficult position. Lékó, most likely in time pressure, overlooked White’s simple reply.
24.Rc6
24...Nf6 25.Rxa6
Carlsen continues, “Now the b5-pawn falls, and White must win. I was somewhat afraid that on a6 the rook would be
out of play, while Black took over the e-file. However, my fears proved groundless.”
25...Bd7 26.Nxb5 Rb8 27.a4 Ng4 28.Bf3 Qh6 29.Qc4 Nxh4 30.Bxg4 Bxg4 31.gxh4 Bf3 32.f5 Qh5 33.Qf4 Bxd5
34.Nxc7 Bb7 35.Rb6 f6 36.Bd4 Qf7 37.Ne6 Rg8 38.Kf2 Rbc8 39.Bc3 Bd5 40.a5 Rc4 41.Nd4 Ba8 42.Qxd6 Qh5
43.Qf4 Rcc8 44.Rbe6 1-0
18...Rg8
With this extremely modest move, Black renders moot all ideas of exploiting the rook’s unfortunate position on h8.
With the rook on g8 now, putting the bishop on f4 is a shot into the air (see the game), and it is White who must
demonstrate his right to exist.
As computer analysis showed, in the diagram position White does have such a right: 19.f4 with good compensation for
the sacrificed pawn – a complex position, close to dynamic balance.
However, Hellers couldn’t find this pawn thrust, and put his bishop on f4 instead.
19.Bf4
Weak.
19...Nbd7
20.Qd2 Bb4
Punishment!
White’s in shock.
22...Nxh5
and Black successfully exploited his material advantage. Hellers resigned on move 47.
Is that all? No! An afterword, important for practical chessplayers, is yet to come.
Black’s position became denser after 18...Rg8 – the rook grew closer to the king. The “traditional” density of packing
for Black’s position for all the pieces was 13/24, where 24 is the area of the “minimal” rectangle a8-a6-h6-h8. But
would it perhaps have been more accurate to replace the number 24 with the figure 20 = 6 + 7 + 7, where 6 is the
distance between the king on c8 and the rook on h8, the first 7 is the distance between the a7- and g7-pawns, and the
second 7 is the distance between the b6-knight and the h6-pawn? The density of packing of Black’s position then would
be 13/20.
A thousand questions... Chess is limitless.
For now, one thing is clear to us: 13/19 > 13/20, where 19 = 5 + 7 + 7 – the rook has made one step in the direction of
the king.
A final query, out of the thousand unanswered ones: is there a kernel of truth in my ruminations?
7...h5
8.Bd3
I will say here that this move is good, despite the fact that it brazenly ignores Black’s counterplay. To me, it seems to be
not at all inferior to the alternative, 8.h4.
Its pluses are obvious: White adds significantly to his second and fifth parameters (“t” and ∆(move)). The drawback of
this bishop move, however, is the opponent’s counterplay.
8.h4 is good, too, and its positives are more than obvious. White not only increases the ∆(move) of his own position, but
also – and this is most important! – nips Black’s counterplay in the bud. The drawback is stagnation in the second
parameter of the position; compare the mobility of the white pieces before and after 8.h4.
Taimanov (White) may have rejected the move because he feared the simple knight maneuver 8...Nh7 … 9...Bf6 and
10...Bxh4:
8...h4
Naturally!
9.Nf1
This move I would argue with: 9.Nge2 looks stronger. Now, hyper-aggressive, unprepared, reckless moves such as
9...h3 10.g3 Bg4 (… 11...Bf3) don’t work. White is saved from a strategic bind on the kingside by the “lucky” check
11.Qa4+:
9...Nh5
10.g3
The knight’s back-and-forth motion leaves an impression. Ivkov intends to do what Taimanov has allowed him to do,
namely 12...Bh3 and 18...Ng4 – see later on – with permanent pressure on the h2-pawn. Outstanding!
11.Bg5
This is wrong – White consistently increases the mobility of his pieces at the cost of their compactness. 11.Rg1 was
better – considerations of compactness predominate over the “t” parameter. The idea is, after 11...hxg3, to play 12.hxg3
and not 12.fxg3 as in the game. Most complicated strategic play, “equalizing” with three unknowns – parameters “t”,
∆k, and ∆(move).
Remember, for the hundredth time (or maybe the thousandth): with “Capablanca,” compactness considerations
predominate. Especially with the left-handed Capablanca: White has something to protect.
Taimanov has taken the first step into the abyss...
11...hxg3 12.fxg3
It was probably better to take back on g3 with the knight, increasing the “t” parameter, since after either pawn recapture
White stands equally badly in the ∆k parameter. And all this with a minus in the ∆(move) parameter for Black.
An important moment: holes have appeared in White’s position, and Black has every reason to release the spring.
12...Bh3
The dream has come to pass. Black problems with his piece arrangement are over.
He can play this way or some other way. Black’s hands are freed – he has the strategic initiative. White is a little worse,
and must stay alert!
15.Kb1 N8h7 16.Bh6
A clear strategic blunder (exchanges are contraindicated for White; 16.Bh4 was better) in adition to a minor tactical
oversight.
16...Kf8
Neither player sees the intricate maneuver of the light-squared bishop after the primitive exchange on h6! Here is
roughly the variation indicated by Rybka: 16...Bxh6 17.Qxh6 Bg2 18.Rg1 Bf3 19.Rd2 Ng4 20.Qg7 Qf6, with a clear
edge for Black.
17.Bxg7+
Wasn’t it better to retreat this bishop to e3? On the other hand, after 17.Be3 Ng4, White is worse.
17...Kxg7
Black’s superiority in the fourth factor is threatening. Tough times for White...
18.Ne3 Ng4
17.Nf3
Petrosian: “If the pawn were on h7, 17...Ng6 would be a good move. But now, after 18.Be3, Black has no satisfactory f-
pawn advance that doesn’t lose a piece. Therefore, the normal move is
17...Qf6
as my opponent played.”
18.Qd2
Where does the queen stand better – here, or on c2? Petrosian has no comment on his last move.
18...Rae8
Rybka prefers 18...Nf3+. Remember, Black is more compact – which means that he is entitled to make an exchange
first.
19.Nxe5
This, shall we say, debatable strategic exchange (Black is more compact!) was conceived by Petrosian back on move 17.
Otherwise, it seems to me, he would have played, not 17.Nf3, but 17.Kh1.
19...dxe5 20.Be3
“One of the most complicated strategic problems, which has occupied more than one generation of masters, is the
problem of hanging pawns. As a rule, they come about (for White as well as Black) in closed openings, primarily in the
Queen’s Gambit. Those who love hanging pawns expect to utilize their dynamic power. Other chessplayers are
prepared to show that hanging pawns have much greater problems than advantages. And of course, there are specialists
who accurately sense every nuance of such positions and successfully play both with and against hanging pawns...
“Later on in the game, we have uninteresting-looking positional maneuvering. So as not to fall asleep from boredom
while examining moves 21 through 32, one should keep in mind the following:
“Hanging pawns are good when they control important central squares, give the possibility of developing the forces
under their protection, then threaten – by the advance of one of the pawns (or both, if the opponent and the situation
permit it) – to destroy the enemy battle formation. While this is happening, one must take into account that the main
positional method for combating hanging pawns is a direct attack on them – which, however, is often only a decoy
aimed at inducing one of the pawns into advancing so that an enemy piece can penetrate and set itself up comfortably.
“This game is a characteristic demonstration of the way to play against hanging pawns” (Petrosian).
Please memorize these words. They have wisdom in them... Petrosian is our mentor!
20...b6 21.Bh5
Commenting on this game, the ninth World Champion acknowledged that he considered this move (most likely,
incorrectly!) a wasted tempo.
21...Re7 22.Bd1
In order to play 23.Bc2, “with a direct attack” (Petrosian’s phrase) on the hanging f5-pawn . With an attack, which “...is
only a decoy aimed at inducing one of the pawns into advancing.”
22.g3, recommended by Rybka, is probably stronger than the game move. Now the tempting 22...Nb7 … ...c7-c5, ...Nd6
is not possible, as 23.Ne4! wins immediately.
The h5-bishop would restrict Black’s possibilities on the kingside in the strongest way possible, whereas the attack on
the target at f5 could wait a bit...
22...Qd6
A natural reaction to the “counterintuitive” stance on the f-file – oh, that “awful” rook on f1! On the other hand, I’m not
sure that Suetin’s move was the strongest... Chess is a very complex intellectual game. and this encounter between
Petrosian and Suetin is a clear example of that. Such games are a genuine nightmare (or just a sweet horror!) for
analysts.
I spent 15-20 hours on this game and wrote out a bunch of computer variations, ten pages of them – over a hundred
possible variations at least 14-15 moves deep. And what did that get me?
Sometimes I feel that I don’t understand a thing...
23.Bc2 Ref7
Of course! The greater the local density of packing in the immediate vicinity of the king, the easier it is to defend.
24.Kh1
A question with no answer: didn’t Petrosian here regret choosing 17.Nf3 over 17.Kh1? Wasn’t he too hasty?
24...Ra8
A mistake! Suetin misses a most curious redeployment: 24...Nb7, ...c7-c5(!), ...Qd6-f6(g6, e7), ...Nb7-d6.
It’s interesting that this setup was also overlooked by Petrosian (in his pre-computer age analysis). Even more
interesting is that Kasparov himself (aided by a computer) didn’t notice it! Why?
Probably because both man and machine found it very hard to find ...c7-c5 – a subconscious fear of the protected passed
pawn. Am I right?
I dare Rybka – I move 24...Nb7. The engine replies 25.b3, expecting 25...Nc5. Amazing!
My own move is 25...c5. I wait...
The reply: 26.Rf3 Qf6 27.Qe2 (27.Ne4 Qh4=) 27...f4 28.Qd3 Bh8² (0.63) at 14 ply (half-moves).
Again, my own move involves waiting with 27...Nd6.
First line: 28.Bc1 Qd8 29.Bb2 Bf6² (0.63), 16 ply.
I don’t believe that the value of ² is so great! The computer can evaluate a position incorrectly, and I’m glad for that
here. In short, the position after 25...c5 and 27...Nd6 is one I like very much. It is pretty! Rybka here is needed to make
sure I’m not overlooking some clever tactical resource.
That’s my dream... Fall in love along with me!
25.Re2
White intends to put this rook on f2 and then bring the queen to h5.
25...Qf8
26.Ref2 Nb7
27.Qe2
Too direct. 27.Nb5 looks better: 27...Rc8 (27...Bxb5 28.Rxf5+-) 28.g3 … Qd2-e2-h5 ±.
27...Nd6 28.c5
Petrosian and Kasparov both give this move an exclamation point. Now “...the number of weaknesses in the enemy
camp begins to exceed the acceptable norm” (Kasparov).
“An incomprehensible maneuver which takes his queen away from the major field of action. 30...e4 was correct, after
which White could not play 31.Nxe4, because of 31...Bb5” (Petrosian). But the e-pawn move, of course, does not end
Black’s suffering: after 31.Qe3 … 32.Nc3-e2-f4 (indicated by Kasparov), White’s advantage is more than evident.
31.Qh5 Qa6
The final step over the edge. Now Black simply must lose – assuming, of course, that White plays perfect chess...
31...f4 (Rybka) leaves Black some small hope for survival. I will not bore you with reams of analysis. Let me say only
that Black is in a bad way, but he’s not losing immediately.
Whereas now...
32.g4
A crushing blow, after which Black’s defense shatters into little pieces, as White enjoys an overwhelming force
superiority in the attack zone.
White engaged the Tal Algorithm on the proper basis.
32...f4
Petrosian: “Forced, but now Black’s threatening-looking passed pawns are easily blockaded...”
33.Re1
Alas, Petrosian is mistaken! The rook’s place is on g1. But not right away: after the preliminary, and winning, 33.g5
(Rybka), White is obligated to attack material targets, and not empty squares in his own half of the board. “Tal” is
“Tal.”
Later on, in Black’s time pressure, both sides did not play their best, to put it mildly. So therefore I decline to present the
rest of the game, such an interesting one to this point. Suetin resigned on move 40.
The final words by Kasparov: “Despite the mutual time-trouble mistakes, this is a very instructive game. It signifies the
triumph for the basic idea of... deep prophylaxis, based on the effect of long-term factors of the position against
temporary, petering out dynamics...”
What do Games 107 and 108 have in common?
Answer: the strategic struggle for control of occupied and vacant squares near the demarcation line.
The Capablanca Algorithm – both its left and its right sides in action. In this, it is important to note its starting point at
1.00.
One more thing: there is a sharp confrontation over the right to dominate. The factor of compactness and the factor of
spatial expansion are in conflict...
We are nearing the finish line – just seven instructive examples to go in this chapter. Complex, even super-complex
examples, with diagnoses ranging from “Tal” to “Petrosian” – the entire chess spectrum is represented.
What do all of these notable games have in common?
Answer: chess beauty! Beauty through paradox.
Beauty is above logic, and synthesis is higher than analysis, since paradox always lies beyond any sort of branches of
any “ultimate” formal/logical system. Render unto God the things that are God’s, and unto Caesar the things that are
Caesar’s...
So close to the end of this chapter, I do not intend to bore you too much, dear reader, with my maxims and
commandments. What will we have next, then?
Answer: a triumphant hymn of primordial chess beauty.
Let it begin!
24...Nc7
Surprisingly, this wins by force. And this, with such a small value in the “t” parameter!
For us mortal chessplayers, it is also curious that also the “super-intelligent” chess engine at first does not particularly
like this move. It prefers the heavyweight maneuverings of the rook, with 24...Ra6 or 24...Ra4, to this piece sacrifice.
25.Qxc7
The only reasonable move, as Black threatened the decisive 25...Rxa3 26.bxa3 Nb5-+.
25...Rc8
Outstanding!
26.Qxe7
26...Nc4
Or 34.Kb1 Qxf8-+.
22...Bb4
With this move, Stein takes the tiger by the tail – playing dangerously! Dangerous for Black – and for White.
After the “correct” and slightly pathetic 22...Nd5 23.Qxe7 Bxe7, the game would most likely have ended in a draw.
While now... now, we will be witnesses to great activity by Black’s king.
23.Bxf6 gxf6
24.Qxh6 f5
25.Ng5
25...Bc3
“This is Stein’s amazingly courageous idea! At first glance, Black doesn’t threaten anything right away, but only
defends. Nevertheless, Black has prepared a hidden threat – 26...Qf6!, warding off White’s attack and keeping all of his
trumps on the d-file. The whole point is that none of the three white pieces piled on that file can move from its spot, so
Black can allow himself the luxury of first defending his king, and only then proceed to the reaping” (Gufeld and
Lazarev, in their book on Leonid Stein).
26.h4
A clear error, probably rooted, not in chess, but purely in psychology – Bisguier overestimates the merits of White’s
position. Had he realized that his position was not better and that he needed to defend, then he could have found the
draw without any particular difficulty. For this, it would have been necessary to forgo “Tal” in favor of “Petrosian”...
The wide avenue to the draw begins with an attempt to give perpetual check: 26.Qh7+ Kf8 27.Qh6+, etc. And if Black
does not consent to this, then he has only one way to continue the fight: 27...Bg7, but this removes his attack on the d2-
rook – a clear achievement for White!
Moving along: 28.Qh4. This is the only move, since 28.Qh5 is not possible in view of 28...f6, and admittedly this does
not require much mental gymnastics. Here Stein would have had to work hard to find the strongest line: 28...Qc5 …
29...Rxd3-+:
Position after 28...Qc5 (analysis)
In this position, White has a guaranteed draw following either the cautious 29.Qg3 (Moiseev) or the active 29.Qh5
(Kasparov), as well as after the aggressive 29.g4, not mentioned by anyone.
Possible variations:
a) 29.Qg3 Rxd3 30.Nh7+ Kg8 31.Nf6+=;
b) 29.Qh5 Qb4 30.Qe2 e5 31.e4 Bh6 32.Qh5 Bxg5 33.Qh8+ Ke7 34.Qxe5+ Kf8 35.Qh8+=;
c) 29.g4 (… 30.Nh7+ Kg8 31.g5 … 32.Nf6+ with attack) 29...Qb4 30.h3 Rd5 31.a3 Qxb3 32.Bc2 (if 32.Nh7+ Kg8
33.g5, then 33...Rxd3 34.Nf6+ Bxf6 35.gxf6 Qxd1+ 36.Rxd1 Rxd1+ 37.Kh2 Bc6-+) 32...Qxa3 33.Rxd5 Rxd5 34.Rxd5
exd5 35.Bxf5, with an approximately equal position.
One gets the impression that after 29.g4 (variation c), it is Black who needs to play accurately in order to maintain
equality.
Leaving these interesting variations behind, we return to the game.
26...Qf6
A forceful punishment for White’s overextending his position (with 26.h4)! As an uncomplicated analysis shows, White
now loses in all variations.
“More elegant than the crude 28...Ke7” (Kasparov). As we can see, the esthetic component predominates... And for
Stein, and for Kasparov, and, you can be sure, for many, many other gifted or simply strong chessplayers.
29.Nh7+
Why didn’t the monster gobble Stein when he tweaked the tiger’s tail? Wasn’t it because Stein was the compact one,
more densely “packed”?
30.g4
An amazing move for its beauty, and paradoxical to boot! White increases his ∆(move) and prepares a safe haven for the
king. And this while having a “Tal” in the middlegame!
And one more thing: there’s a... Zugzwang? On the board! Could that be?
30...Rb7
Knight checks are pointless: after 30...Nc2+ 31.Kf3 Nd4+ 32.Kg2, White wins. And 30...Nc6 31.Qe6+ Kb7
32.Qxg8+- is quite bad for Black. Nor does 30...b5 help: after 31.Qxc5+ Nc6 32.Rd1, Black can resign with a clear
conscience.
One more problem – the king is paralyzed (30...Kc7 31.Qxg8+-). Add to that problem two more great ones – the
immobile queen and the poor rook on g8.
Black has only a-pawn moves left... Here’s a variation by Krasenkow: 30...a6 31.f3 a5 32.Kf2 a4 33.Kg2 a3 34.Kf2:
Position after 34.Kf2 (analysis)
Zugzwang. Verify this – examine all of Black’s moves and find their immediate refutation. Zugzwang in the
middlegame is a very rare event!
Let’s return to the game and its spectacular finish.
31.Qxg8
This wins.
31...Qxg8 32.Rh8 Rxe7 33.Rxg8+ Kb7 34.Rxg6 Rf7 35.Rh3 Nc2+ 36.Ke2 Nd4+ 37.Kf1 1-0
No. 112: Atalik – Sax
Szeged 1997
17.dxe5
An opening novelty, breaking up what appears to be a position that has been studied inside and out. Now comes
independent play from GM Sax.
17...Be6
18.Nf3
The point of White’s idea. I add: the computer got interested in this move only at a depth of 14 ply. Before that, its top
two lines were 20.Qf4 and 20.Qf3.
20...Rxf7
21.Bxe6
The critical position. According to analysis, Black is much worse. Question: is there a saving line?
A careful answer (and, of course, one which by no means aims at absolute truth): more than likely, Black’s position is
indefensible.
Commenting in Chess Informant on this notable game, GM Suat Atalik risked bringing under his readers’ judgment one
lengthy – it would be a sin to hide it – and debatable variation. Here it is: 21...Rxf2!? 22.Kxf2 Rf8+ 23.Kg3 Qe1+
24.Kh3 Kg7 25.Bd7 Kh6 26.e6 Rf5 27.Bf6!! Rh5+ (27...Rxf6 28.e7+-) 28.Kg4 Rf5 29.e7 Qd1+ 30.Kh3! Qd7
(30...Rh5+ 31.Kg3 Qe1+ 32.Kf4 Qf1+ [32...Qf2+ 33.Qf3 Rh4+ 34.Bg4!; 32...g5+ 33.Ke5 Qc3+ 34.Qd4 g4+
35.Kf4+-; 32...Rh4+ 33.Bxh4 Qxh4+ 34.Kf3+-] 33.Qf3+-) 31.e8Q Rf3+ (31...Rh5+ 32.Kg3+-) 32.Kh4+-:
21...Rg7
22.Bf7
This position, it seems to me, is much simpler than that following 21...Rxf2. A sample variation (Atalik + Rybka + my
imagination):
22...Qd1 23.e6 Qd6+ 24.f4 (24.Be5 Qe7 25.Bxa1+-; 24.Kh3+-) 24...g5 25.Kg4 Qd1+ (25...gxf4 26.Kh5 Qc5+
27.Qe5+-) 26.Kh3 (26.Kf5 Qd6±) 26...g4+ 27.Kh4 g3 28.h3 Qd6 29.Kh5 Rag8 30.Kh6 Qf8 31.Qxh7#:
Impressive, huh?
If you don’t like that long variation (if it seems unpersuasive to you), then try to improve on Black’s play. I couldn’t do
it...
In the game, Sax played a weaker line and lost without a struggle.
24.Qd4
26.Qd7+ 1-0
Sax resigned, since after 26...Kf8 27.Bf6 there is no defense against mate.
Ineffable beauty...
16.f5
“Of course, this piece sacrifice is of an intuitive nature, as all of the variations could not be calculated over the board.
Black’s idea is revealed after 16.fxe5 Bxe5 17.Bh6 (on 17.Nf3, 17...Bg4 is not bad) 17...Bxh2+ 18.Kh1 Be5! 19.Bxf8
Rxf8 – despite the loss of the exchange, Black has full positional compensation” (Sakaev, Shakhmatny Peterburg, Nos.
4-5, 2003).
The text move was not a virtual, but a real sacrifice. For Sakaev, we will see that the initiative was worth more than
material.
The two first lines of Rybka (at a depth of 13 ply):
1) 16.fxe5 Bxe5 17.Nf3 Bg4 18.Qc2 Bxf3 19.gxf3;
2) 16.f5 exd4 17.cxd4 gxf5 18.exf5 Bf6 19.Rb6 (at this depth of calculation, the computer could not give a proper
evaluation of 19.Qh5 – see the analysis position in the diagram following the next one).
16...exd4 17.cxd4
Position after 17.cxd4
The parameters of the position from Black’s standpoint are: m > 1, t = 43/46 = ~0.93, approximate parity or slight
inferiority in the safety factor, ∆k > 0, and ∆(17.cxd4) = ~-0.54.
Diagnosis for Black: the expanded version of the CP Algorithm. That means that the diagnosis for White, anti-
symmetrical to Black’s, is the TC Algorithm.
The zone of uncertainty has widened (compare this with the assessment of the position after 16...e5). It may even have
stepped over the boundary of the Capablanca Safety Zone and taken up the part of the spectrum from t = ~1.0 to t =
~1.5. Of course, in our case, we can only be talking about approximate boundaries for the zone of uncertainty.
Analyzing the position after 17.cxd4, I flesh out Sakaev’s analysis, even if only a little. I did, however, find a major hole
in Krasenkow’s Megabase analysis. And here is my careful verdict on the study position: the piece sacrifice is correct,
and 16.f5 is, most likely, the strongest move.
In GM Sakaev’s opinion, the best defense for Black would have been 17...gxf5 18.exf5 Bf6. However, one may argue
with this, since White does have a very strong reply which gives him the advantage: 19.Qh5 (indicated by Rybka after
very heavy thinking):
Position after 19.Qh5 (analysis)
Out of the many dozens – even hundreds – of possible variations in this highly complicated position, I have selected
two. One of them is long (meaning: arguable) but carefully selected:
a) 19...Rfe8 20.Rf4 (20.Rf3 Rxe3! 21.Rxe3 Bxd4 22.Qg4+ Bg7 23.Rc1 Nc6, “and in the resulting very complex
position, Black is no worse” – Sakaev) 20...Rxe3 21.Rg4+ Kh8 22.Qh6 Qd8 23.Qxe3 Bb5 24.Rd1 Nc4 25.Bxc4 Bxc4
26.d5±;
b) 19...Rae8 20.Rf4 Bb5 (20...Rxe3 21.Rg4+ Kh8 22.Qh6 Qd8 23.Rh4 Bxh4 24.f6+-) 21.Rf3 Rxe3 (21...Bxd3
22.Rh3+-) 22.Rxe3 Bxd3 (22...Bc6 23.Qh6 Bg7 24.Qh4 f6 25.Rbe1±) 23.Rxd3 Qf4 24.d5 Bd4+ 25.Kh1 Bxc5 26.d6
Nc6 27.d7±.
Our conclusion: Black is worse after 17...gxf5 18.exf5 Bf6 19.Qh5.
So perhaps it would be better to play, not 18...Bf6, but 18...f6 instead? Let’s take a look:
Position after 18...f6 (analysis)
Obviously, 18...f6 instantly rids Black of all unpleasantness connected with an attack on his king. But oh, that bishop on
g7!
Nevertheless, I like the pawn move better than 18...Bf6 because it is safer, as analysis shows. Some likely variations:
19.Bf4 Qd8 20.d5 Rc8 21.Qc2, and then:
a) 21...Be8 22.d6 Bf7 23.Rfe1 Re8 24.Be4 Qd7 25.Bd2 Nc6 26.Rb6 (… 27.Qb1 or 27.Reb1±/², with powerful pressure
on the b7-pawn);
b) 21...b5 22.a4 Be8 23.Be4 Nc4 24.c6 Bf7 25.axb5 axb5 26.Qf2 Re8 27.Bf3 Nd6 28.Bxd6 Qxd6 29.Rxb5².
Please verify this for yourself!
Let’s return to the game we have almost forgotten – that is, to the position after 17.cxd4.
17...Rfe8 18.f6
18...Bf8
The natural retreat. However, oddly enough, the “unnatural” retreat to h8 also has a claim on our conscisousness: after
18...Bh8, Black – even in the variations – never does get mated on g7, as White cannot trade off the dark-squared
bishops. Possible variations include:
18...Bh8 19.Rb6 Bb5 20.Bf4, and now:
a) 20...Qc8 21.Bd6 Re6 22.e5 Nc4 23.Bxc4 Bxc4 24.Rf4 (Sakaev’s recommendation of 24.Qf3 is suspect: 24...Bxf1
25.Kxf1 Qe8 26.Rxb7 Rc8 … 27...Bxf6 with powerful counterplay – Black is better!) 24...Bd5 25.Qd2 Qe8 26.Rb2 (…
h2-h3, Kh2, Rh4, Qh6+-) and Black is defenseless;
b) 20...Qd8 21.Rd6 Qc8 22.a4 Bxd3 (22...Bc4 23.e5 b5 24.cxb6 Bxd3 25.Qxd3 Nb7 26.Rd5±) 23.Qxd3 Re6 24.e5
Rxd6 25.cxd6 h5 26.Rc1 Qg4 27.Bg3 Qg5 28.Rc7±/².
As we can see, in that last variation Black manages to hold the position somehow. Or, at least, he is not losing by force.
19.Qc1
“Intending to use the h6 square to give mate. 19.Qd2, with the same goal in mind, would not work because of 19...Bb5
20.Bh6 Bxc5! 21.dxc5 Red8!, and the roles change – Black becomes the attacker” (Sakaev).
The position after 19.Qc1 (see diagram) is also unspeakably complex; and I admit that I could not exhaust it totally. It’s
likely a tiny bit better for White, but... But, like the neutrons in an atom bomb, the variations deserving attention flare
out exponentially after every move, and I lose control of the whole process...
Some variations I like:
a) 19...Bb5 20.Bxb5 axb5 21.Bh6 Bxc5 (21...Rxe4 22.Bxf8 Kxf8 23.Qh6+ Ke8 24.Rfe1 Qc6 25.Qh4! Rxe1+ 26.Rxe1+
Kd8 27.Qxh7+-) 22.Qxc5!? (22.dxc5 Rxe4 23.Rxb5 Nc4) 22...Qxc5 23.dxc5 Rxe4 24.Rxb5 – Krasenkow’s variation.
My evaluation of the final position is ±.
b) 19...Kh8 20.Bh6 Qd8 21.Bxf8 Rxf8 22.Rb6 Rg8 23.d5 Bb5 24.Bxb5 axb5 25.Rxb5±.
c) 19...Rac8 20.Bf4 (20.Bh6 Bxc5²) 20...Qc6 21.Bh6 Bxc5 22.dxc5 Qxc5+ 23.Qxc5 Rxc5 24.Rfc1².
The unanswered question is: is 19...Rac8 the best move here?
GM Belov’s choice was likely no worse. Let me put it more delicately: it’s a worthy challenger to the rook move!
19...Nc6
Coming closer to the king. Now the tempting 20.Bh6 doesn’t work, as after 20...Nxd4 21.Bxf8 Rxf8 22.Qh6 Ne6 Black
fends off the attack without difficulty. Among other things, what saves him is that he can take the c5-pawn with check.
20.Kh1
Prophylaxis. Now the capture on c5 no longer comes with check – super-fine play! However, it is possible that White
would have had better chances for success with the somewhat rough 20.Rb6:
Position after 20.Rb6 (analysis)
Wandering along with Rybka through the impenetrable jungles of variations, I select a trail – one of a thousand possible
ones. Here it is:
20...Rad8 21.Bh6 (21.Bf4 Ne5 22.Qe3 Qc8 23.Bxe5 Rxe5 24.c6 Rh5 25.cxd7 Rxd7, and Black may even be just a bit
better) 21...Bd6 22.Bf4 Bxf4 23.Rxf4 Nxd4 24.Rh4 Nb3 (and what about 24...Ne2+?) 25.axb3 (try out 25.Rxb3 Re6
26.Qg5) 25...Qe5 26.h3 h5 27.Qh6 Qxc5+ 28.Kh2 Qf8 29.Rxh5 Qxh6 30.Rxh6 Re6:
20...h5
Sakaev thinks that this is the only move. However, Krasenkow disagrees, having discovered a paradoxical-looking
bishop move: 20...Bg4!?, and then 21.h3 (see diagram below; inferior is 21.Bh6 Nxd4 22.Bxf8 Rxf8 23.Qh6 Ne6
24.Bc4 b5 25.Bd5 Ra7 26.Rbe1 Qxc5 27.h3 Bh5 28.g4 Bxg4 29.hxg4 Qc3) 21...Bc8!? 22.Bh6 Nxd4 23.Bxf8 Rxf8
24.Qh6 Ne6 25.Rb6!.
Now we see the point of 20...Bg4: after 25.Bxe6 (instead of 25.Bd5) 25...fxe6, Black’s queen safely defends the
Achilles’ heel of the position – the g7 square – from c7. Bravo, Krasenkow! Bravo, and... oh, the horror!
21.Bf4
White also has the upper hand after the quieter 21.Rb6 (Rybka).
21...Qd8
Or 21...Qc8 22.Bd6².
22.Bd6 Kh7
Inaccurate, to put it mildly. 22...Bxd6 23.cxd6 Kh7 was better.
23.Qe3
Returning the favor! 23.e5 was stronger, and if 23...Bxd6, then 24.exd6, with a large and almost decisive advantage.
23...Bxd6
“On 23...Bh6 there would follow simply 24.Qf2, with e4-e5 to come” (Sakaev).
24.cxd6
24...Bc8
This loses, whereas another move, 24...Qb8, offers hope. Here’s Rybka’s variation: 25.e5 Qxd6 26.Qf3 (26.Rxb7²)
26...Nxe5 27.Qxh5+ Kg8 28.dxe5 Rxe5 29.Qh6 Qf8 30.Qxf8+ (30.Qh4 Rh5 31.Qg3 Qb8²/=) 30...Rxf8 31.Rxb7 Be6.
White is better, but a draw is the most likely outcome.
25.e5
25...Qxd6
26.Bxg6+
29...Qe7
“For instance, 29...Kg7 could be met by 30.Rbf1 Be6 31.d5! Bxd5 32.e6” (Sakaev).
One of the most complex games we have seen so far. But wait, we still have Polugaevsky – Torre (Game 120), an
analyst’s nightmare!
24.Rxd4 cxd4
This move is refuted in a hardly obvious way. Commenting on this game for Shakhamatny Peterburg, No. 1, 1999, GM
S. Ivanov rejected this move, as did GM Stohl for Megabase. Their analyses were independent of one another, varying
in the details.
At least three other moves were stronger than the text: 24...g5 (Rybka), 24...Bxd5, and 24...Kb6. Both Ivanov and Stohl
consider the king move strongest in this position. No argument there!
Only like this! And first of all because the natural 25.Qxd4+ doesn’t work, since after 25...Qb6 26.Re7+ Nd7 27.Rxd7+
Rxd7 28.Qxh8 now it’s Black who wins – 28...Qg1+ 29.Ka2 Bxd5+ 30.c4 (30.b3 Rc7-+) 30...bxc4-+ (Rybka).
25...Kb6
“Black must go forward, as 25...Qxe7 loses: 26.Qxd4+ Kb8 27.Qb6+ Bb7 28.Nc6+ Ka8 29.Qa7# or 25...Kb8 26.Qxd4
Nd7 (26...Rd7 27.Rxd7 Nxd7 28.Qxh8++-) 27.Bxd7...” (S. Ivanov). After 27...Bxd5 28.c4! (S. Ivanov, Stohl), Black is
defenseless. Check it out!
26.Qxd4+ Ka5
27.b4+ Ka4
28.Qc3
This probably wins. However, Stohl indicated a slightly stronger move, 28.Ra7. Stohl’s variations:
a) 28...Bxd5 29.Qc3 Rhe8 30.Kb2 Re2 31.Qc7+-;
b) 28...Nxd5 29.Rxa6+ Qxa6 30.Qb2 Nc3+ 31.Qxc3 Bd5 32.Kb2 Qe6 33.Bxe6 fxe6 34.Qb3+ Bxb3 35.cxb3#;
c) 28...Bb7 29.Rxb7,
c1) 29...Nxd5 30.Bd7 (… 31.Bb5++-) 30...Rxd7 (30...Ra8 31.Bxb5+ axb5 32.Ra7+ Qa6 33.Qxd5+-);
c2) 29...Qxd5 30.Rb6 a5 (30...Ra8 31.Qxf6 a5 32.Bf1 … Qc3, Kb2, Qb3++-) 31.Ra6 Ra8 32.Qe3 Rxa6 33.Kb2 axb4
34.axb4:
Position after 34.axb4 (analysis)
White wins!
And what happened in the game?
28...Qxd5
29.Ra7 Bb7
29...Rd6 was bad: 30.Kb2, and Black can’t play 30...Qd4 because of mate on a6.
30.Rxb7
30...Qc4
Ivanov and Stohl both view this move with suspicion, preferring 30...Rhe8. True, this doesn’t stop Stohl, after 31.Rb6
Ra8 32.Bf1, from burdening Black with imminent defeat – that is, “+-”.
Is that so? More than likely, as 32...Rec8 33.Qxc8 Qd1+ 34.Ka2 Qd5+ 35.Bc4 Qxc4+ 36.Qxc4 bxc4 37.Rxf6 (Rybka)
yields a lost ending for Black. Try it!
31.Qxf6 Kxa3
“The black king goes bravely forward – to his doom. Now Black loses by force. After 31...Rd1+ 32.Kb2, he could trade
queens: 32...Qd4+ 33.Qxd4 Rxd4, but this would not change his fate: 34.Rxf7 Rd6 (34...a5 35.Ra7+-; 34...Re8 35.Rf6
Ra8 36.Be6+-) 35.Re7! … Be6+-.
“And in the case of 32...Ra8 33.Qb6 (33.Rxf7 Rd2) 33...Qd4+ (33...a5 34.Bd7 Rd5 35.Qe3+-) 34.Qxd4 Rxd4 35.Rxf7
a5 36.Be6 axb4 37.Bb3+ Ka5 38.axb4+ Kb6 (38...Rxb4 39.c3+-) 39.Rxh7+-, Black cannot hold the position” (S.
Ivanov).
32.Qxa6+ Kxb4
33.c3+
Remarkable.
33...Kxc3
Or 33...Qxc3 34.Qxb5+ Ka3 35.Ra7++-; 33...Kc5 34.Rc7++- (S. Ivanov). Or 33...Kb3 34.Qa2+ Kxc3 35.Qb2+ Kd3
36.Re7!+- (Stohl).
34.Qa1+ Kd2
35.Qb2+ Kd1
Black loses after 35...Ke1 36.Re7+ Kd1 37.Bf1 (Stohl). Let’s play out this variation: 37...Rd2 38.Be2+ Rxe2 39.Rxe2
Qd3+ 40.Rc2++-.
36.Bf1 Rd2
39.Qxh8 Rd3 40.Qa8 c3 41.Qa4+ Ke1 42.f4 f5 43.Kc1 Rd2 44.Qa7 1-0
The black king, having reached the first rank, is forced to sign articles of unconditional surrender.
20.Qc3
The right-hand – that is, the “Tal” branch of the TC Algorithm: White sacrifices a pawn for an attack on the king. The
alternative is the “Capablancan” 20.Rd2, which leads to the trade of the “bad” d-pawn for the b-pawn. Rybka gives the
variation 20...Nxd4 21.Bxd4 Rxd4 22.Rxd4 Qxd4 23.Bb3 b5 24.Qb7 Rf8 25.Qxb5².
I was hoping for more. More precisely, I didn’t see any moves like 20.Rd2 or 20.Qe3 (Rybka’s second line), because I
was consumed by an irrepressible feeling of chess aggression. Indignantly, after rejecting the cowardly 20.d5 due to
20...Nd4³, I cast myself into the depths of irrational complications, without any inner resistance. And Caissa turned
away from Korchnoi and smiled broadly on me!
20...Nxd4
The threat is 21.d5+-, so Black is forced to capture the pawn, and with the knight in particular, as otherwise (after
20...Qxd4 or 20...Rxd4) 21.Qg3 wins.
21.Re7
21...Rxe7
“21...Kf8, driving out the dangerous white rook, deserved attention. For example, 22.Rxd7 Rxd7 23.Qc8+ Rd8 24.Qc4
Rd7, and Black holds on (S. Ivanov, A. Kentler, V. Faibisovich, B. Khropov, Shakhmatnaya Letopis Peterburga 1900-
2005. Chempionati goroda).”
22.Rxe7 Nf5
Korchnoi’s laconic annotations: 22...Ne6 23.Qb3±; 22...Qd6 23.Qe1!± (see Chess Informant No. 15).
23.Rxf7
The sensible 23.Re1, cutting off my opponent’s terrible counterattack, would have amounted to moral capitulation.
However, I saw no refutation of the game move, and so – this is important! – lost my instinct for self-preservation.
23...Rd1+ 24.Kh2
24...Qd6+
Now 25.Qe5 is bad (and I forgot to mention it in my annotations) due to 25...Ng4+, with mate next. However...
26...Nh4+
“A move like this is easy to find, since it’s strictly necessary. The alternative is to be checkmated: 29.Kxg4? Rg1+
30.Kh5 g6+ 31.Kh6 Qxh4#” (Shakhmatnaya Letopis Peterburga).
29...Qe2+
Loses by force. The rook check 29...Re1+ left him some chances to draw, with correct and “non-human” play...
Selected variations after 29...Re1+ 30.Kd5:
Position after 30.Kd5 (analysis)
a) 30...Qxf7 31.Kd6 (31.Kc5 Re5+³) 31...Rd1+ 32.Kc5 b6+ 33.Kb4 a5+ 34.Ka4 b5+ 35.Kxa5+-:
34...Qxb2
35.Rxg7+
Black resigned (35...Kf8 36.Rg8#). 1-0
As the players in the championship joked, the queen with her own body prevented the king’s access to e7.
Victory!
The crazy white king, walking around the whole board (and that with queens, rooks, and minor pieces still present!),
was protected by Caissa. Why?
There’s no rational explanation for it, and there probably cannot be.
Chapter 4
White to move
1.Nf5 1-0
21...Rf7
Alas, this loses a piece by force – Black gets careless and overlooks a two-move maneuver by the white knight.
After 21...Rc8 22.Rb7 Rf7, analysis shows that Black has no problems – we have dynamic balance.
And on 21...b6, Black probably is even slightly better, and it is White who must then find a narrow path to the draw. For
example, 22.Ra1 Nb4 23.Rc1 (23.Raxa7 Rxa7 24.Rxa7 Nc6³) 23...Rf7 24.Rxf7 Kxf7 25.Rc7+ Kg8 26.Ne5 (Rybka).
Rybka sees no advantage for Black, either, after the cool 22.h3. A draw would have been the proper outcome of the
struggle.
24...Rc8
Nor does 24...a5 help: after 25.Ra1 Nb4 26.Nxb4 he can’t play 26...axb4 – that rook on a8!
White is winning – the knight is stronger than the two pawns, while the rook on the seventh rank can be driven away
easily.
We have exhausted our theme for this game – the theme of the second-pass evaluation in determining the genuine
diagnosis of a position. So I will not comment on it further. I’ll only point out the winner’s flawless technique and the
irrelevant (to the outcome) mistake by Black on move 35.
28.Kf1 a6 29.Nf3 Rc2 30.Nd4 Rc4 31.Nb3 g5 32.h3 h5 33.Nd2 Rb4 34.b3 e5 35.Rc2 a5 36.Rc5 Rd4 37.Ke2 b6
38.Rc6+ Kf5 39.Rxb6 a4 40.Nf1 e4 41.Ne3+ Ke5 42.Nc2 1-0
27.Ba6
An open attack on the rook and – this is the main thing! – a prelude to massive trades. In other words, White plays
according to “Tal” and... against “Tal” – a paradox! A reasonable question: is such strange play correct?
Yes. Indeed, it is more than correct: computer analysis shows that Black’s position is most likely lost.
Black intends, after 30...Bb7 and 31...Rc4, to drive the hated knight from d4 – but, no!
30.b3
Bravo! White forces the exchange of his opponent’s last good piece.
The only practical chance – Black activates his bishop at the cost of the b- and d-pawns.
If Black plays passively (36...Kf8 … ...Kf8-e7-d7-c7-b6), then after 37.Ke3 White succeeds, barely (confirm this!) in
marching his king over to the strategically vital square b5. Then White brings the knight to d4, so as to prepare the
“final” f4-f5. An additional resource (let’s say, with the bishop on c8) would be Nd4-f3-g5, and Black is defenseless
since sooner or later he will lose either the f-pawn or the
h-pawn.
The superiority of the “good” knight over the “bad” bishop is obvious – no wonder White played for massive trades!
After –
– the d-pawn fell, and White finally won (not without mutual mistakes) on move 57.
We are keenly aware of the fact that all pieces except the king have a finite chess value. Let’s rank them in a way well
known to us from the time of “Tal”: queen, two rooks, four minor pieces, eight enemy pawns...
Now we fill out this incomplete array of material targets (the occupied squares of the chessboard) with selected vacant
squares. Let these be, for example, the light squares d3, e4, f5, and g4 – see the diagram:
11...c4
Depriving White’s light-squared bishop of the d3 square and limiting its mobility. This restriction strategy will be
continued – see 12...g6 and 13...h5!.
12.f4
12...g6
The cage closes! Now Black has only to prepare calmly and carry out his attack on the king. Success is guaranteed – in
effect, Black has an extra piece!
14.Bh4 Be7
15.Bg5
A crude tactical blunder (I suspect that Yates simply “closed his eyes”). A mistake in a strategically hopeless position.
For us, the balance of the game is insignificant. More exactly, insignificant for our study of the theme of the second pass
in figuring the algorithm. And therefore I, with a clear conscience, make use of the author’s right to “comment without
commentary.”
15...Bxg5 16.fxg5 Nxe5 17.Qe3 0-0 18.Be2 Nc6 19.f4 Ne7 20.Rhg1 Nf5 21.Qf2 Rfe8 22.Rge1 b5 23.Bf1 a5 24.Re5
b4 25.Qe1 Reb8 26.Ka1 Rb6 27.Qd2 Rab8 28.Rb1 Nd6 29.Bg2 Nb5 30.cxb4 c3 31.bxc3 Nxc3 32.Rb3 axb4 33.a3
Ra6 34.Re3 Rba8 35.Rexc3 bxc3 36.Qc1 Qc5 37.Ka2 Qc4 38.Ka1
38...Qxb3 1-0
White resigned, as he gets mated after 39.cxb3 Rxa3+ 40.Kb1 Ra1+ 41.Kc2 R8a2+ 42.Kd1 Rd2+ 43.Ke1 Rxc1#.
Attacking selected vacant squares and playing to restrict the mobility of a piece are the same thing. They’re just
different terms referring to the same chess phenomenon.
The first term comes from “Tal;” the second, from “Petrosian.” Somewhere in the depths of the Capablanca Algorithm,
they joined together and lost their respective identities.
The following game is unique. Why?
First, it’s very complicated – the computer often radically changes its evaluations.
And secondly, it’s a serious test for what we have labeled the “second-pass evaluation.”
I acknowledge that my many hours of computer analysis in this game cannot be labeled comprehensive at all. I often
lost control of the position. I was drowning in a sea of variations – quite a lot of unexpected tactics. And there were
strategic discoveries, too!
So I invite you, dear reader, to be especially careful in your evaluations.
17.h4
Kasparov says (My Great Predecessors, Part III): “Very fine!... Earlier they played the automatic 17.exf8Q+ Kxf8!
18.Rd6 Rb8 19.Be3 Rh5! 20.Be2 Re5 21.Kd2?! b4 22.Nd1 Kg8! (Plachetka – Bagirov, Berlin 1979), or 21.Nd1 Kg8
22.Bf4 Ree8 23.Ne3 Be4 24.f3 Bg6 25.h4 Rb7 26.g4 Na4 (Beliavsky – Bagirov, Moscow 1981), and in either case
Black has good counterplay on the queenside.”
The h-pawn move is in “Tal” style. But at the same time it’s also “Capablanca” and “Petrosian”! Why?
“Tal,” because it is an extended (by two moves) rook sacrifice. Precisely the rook, and not the exchange, as Polugaevsky
declined to play 17.exf8Q+.
“Capablanca,” because it is the beginning (sorry, the continuation!) of a space expansion. To the highly elevated pawns
at e7 and f6, we rush to add the assistance of the h-pawn.
“Petrosian,” because this is where I expected to find it, since this move prepares a cage for Black’s king’s rook. White is
playing to restrict that rook’s mobility.
Returning to the diagnosis. Is it the TCP Algorithm, yes or no?
A favor to ask: don’t be hasty in answering...
17.h4 Bh6
I won’t rule out the possibility that this may be a serious mistake. It seems to me that it would have been better to play
17...b4 18.Rd6 (18.exf8Q+ Kxf8=) 18...Bxe7 19.fxe7 (19.Rxb6 Bd8-+) 19...Nd7 … 20...f6.
I have found no decisive advantage for White after 19...Nd7. Some variations:
1) 20.Nd5 Bxd5 21.Rxd5 f6 22.Be3 Kxe7;
2) 20.Na4 f6 21.Bf4 c3 22.Nxc3 Kxe7;
3) 20.Bh3 f6 21.Bxd7+ Kxe7 22.Rxf6 bxc3 (22...Kxd7 is possible, too) 23.Bf5 cxb2 24.Rb6+ Kf7 25.Bg6+ Kg8
26.Rxb2 Bf3 … ...Bh5-f7.
White is better in all lines – from ² to ±. Try and refute that!
18.f4
The focus of White’s idea is the crippled h8-rook. Polugaevsky is more than outstanding!
“After 19...bxc3 20.Rxb6 cxb2 (if 20...c2 21.Kd2 Kd7 22.Bxc4 Be4, then even the paradoxical-looking 23.Rxa6! wins)
21.Bxc4 followed by Rxb2 and Black loses because of his weakness on f7 and the amazing helplessness of his rooks.
And, after putting his rook on the d-file, White can even trade off the bishops and play a unique ʻthree-rook’ endgame!”
(Polugaevsky)
20.Nd1
Of course, I have no argument whatever with this natural move. And no argument with the previous three half-moves,
either.
20...Bxg5
But I do have questions about this move, because it is not the only possible one. In the diagram position, Black has at
least two interesting alternatives.
Kasparov: “The consequences of 20...Be4!?, which is not mentioned by the commentators, should have been clarified.
For example, 21.Ne3 c3 22.Bxa6 (not 22.bxc3?! bxc3 23.Bxa6 Nd7 24.a4 Rb1+ 25.Kf2 Rb2+) 22...Nd7 23.Nc4 cxb2
24.Nxb2 Bb1 (see next diagram) 25.Rd1! Bxa2 26.Bc4!, and if 26...b3?, then 27.Bb5! Rxb5 28.Nc4 and wins, while if
26...Bxg5 27.fxg5 b3, then 28.Bf1 is good.”
Excellent! Kasparov has fired the starter pistol. He has provoked an analytical tsunami, and I certainly cannot let this
humble reply go unanswered!
I will analyze two episodes, two positions which interest me.
The first is:
Position after 24...Bb1 (analysis)
White indeed wins after 25.Rd1 Bxa2 26.Bc4 b3 27.Bb5 Rxb5 28.Nc4, i.e. 28...Rb6 29.Nxb6 b2 30.Nc8, mating.
It is also the case that after 26...Bxg5 (instead of 26...b3) 27.fxg5 b3, White doesn’t have the “winning” move 28.Bb5 –
28...Rxb5 29.Nc4 Nxf6 30.gxf6 (the bishops are traded off, and White no longer has 30.Bxf6) 30...Rb8, and it’s Black
who wins! So White must limit his ambitions, as Kasparov rightly points out, to “the very good 28.Bf1.” Rybka extends
the variation: 28.Bf1 Rb6 29.Bh3 Nxf6 30.gxf6 Rxf6 31.Rd8+ Ke7 32.Rxh8. White is better here, possibly even ±. But
there is nothing forcing yet!
But the winning follow-up to 24...Bb1 (it seems to me) is not the Kasparovian 25.Rd1, but 25.Nd3. Here’s my main
line: 25.Nd3 (… 26.Nxc5 Nxc5 27.Bb5++-) 25...Bxd3 26.Bxd3 Rb6 27.Rxb6 Nxb6 28.Bb5+ Nd7 29.Kd2 (… Kd3-d6
and Bd7#) 29...Bf8 30.Kd3 Bxe7 31.fxe7 Rg8 32.Ke4 f6 33.Bxf6 Rxg3 34.h5+-. See for yourself!
I can’t avoid making a snarky remark about the bishop maneuver on moves 29 and 30 which saves him from mate on
d7: wouldn’t it have been better to give up his bishop for the e7-pawn (see my note to 17...Bh6)? Then Black would
have had good chances to draw, whereas here we have a catastrophe.
The second position is, if instead of the game move, 20...Bxg5, we were to play 20...Bd5:
Position after 20...Bd5 (analysis)
This position cost me a whole mountain of time...
In his defense, Black wishes to set up the bishop on e6, closer to his king. From here, the bishop would cover d5 and f5
(he has to consider Nd1-e3-d5(f5)). Besides, from d5 (or e6) the bishop looks greedily at that helpless pawn on a2, and
that is, I admit, very disturbing to me. It seems to me that with the disappearance of White’s a- and b-pawns (let’s say,
after ...c4-c3 and ...Bxa2), his dreams of victory would similarly vanish. Fortunately, my fears were not all to be
realized!
Clarity came to me unexpectedly. It arrived the minute I asked myself this simple question: by what algorithm, strictly
speaking, should I be governed? By “Tal,” or “Capablanca,” or “Petrosian”?
I am ashamed to admit that I had given in to my “pro-Petrosian” prejudice, as the board shows all the signs of the Tal
Algorithm (the radical “Tal” or the TCP Algorithm?). I had to attack, and Target No. 1 was the king!
At once, a cascade of beautiful variations came tumbling out as if from the horn of plenty. One of them was 21.Bh3 c3
22.bxc3 Bxa2 23.cxb4 cxb4 24.Ne3 (… Nf5) 24...Bxg5 25.fxg5 Be6 26.Rxe6 fxe6 27.Bxe6:
Position after 27.Bxe6 (analysis)
White wins.
But that’s not all! Instead of 25...Be6, Black has one more resource – 25...Rh7. Black is ready with exceptional joy to
give up his rook for the knight (after Ne3-f5-g7), but now there is 26.Bg2 Be6 27.Nc4!:
21.fxg5 Nd5
Black intends to give up the knight for the tandem of white pawns.
The advanced duo is destroyed. Does Black have any drawing chances now?
To answer that question, we need to carefully investigate the resulting position:
The parameters: m < 1 (true, he could capture the pawn), t = 35/31 = ~1.13, an insignificant “plus” in the safety factor
(adding that we have an endgame on the board!), ∆k < 0, and ∆(23...Kxe7) = ~-0.14.
Diagnosis: the TC Algorithm, or a plain version of the Tal Algorithm.
24.Rf6
“It is much more important to prevent the rook at h8 from coming into play, than to go after the a6-pawn”
(Polugaevsky).
Kasparov continues: 24.Rxa6 Rhe8 25.Rf6 Kf8+. His verdict: “...the activation of his rook would have enabled Black to
save the game.”
Rybka’s line: 26.Kf1 Re7 27.Nf2 Ba8 28.h5 Rbe8 29.Kg1 Rd7 30.g6 (30.h6 Re3) 30...Re3 31.g4 Rf3 32.Rxf7+ Rfxf7
33.Bxf7 (33.gxf7 Bd5) 33...Bd5 34.Bxd5 Rxd5 35.Ne4 Kg7, and Black is not worse. Which means that we have
thoroughly discredited the greedy 24.Rxa6.
Whereas, after the game move, White (as shown by computer analysis) must win.
24...Rhf8
If 24...Rh7, then White wins with 25.Ne3 Be4 26.Rxa6 Rd8 27.Ng4 Rd4 28.b3 Bg6 29.Nf6 Rh8 30.Ke2 Kf8 31.Ke3
Kg7 32.Rc6 Rdd8 33.Rxc5 Ra8 34.Rb5 Rxa2 35.Rxb4.
Of course, one should be suspicious of long variations (even when sanctioned by the latest chess engines)... Consider
them and try to refute them – if you can!
26.h5, declining the pawn, looks simpler: if 26...a5, then 27.Nf5+ Bxf5 28.Rxf5, with a permanent attack against the
target at f7, looks very strong. Here’s a variation from Rybka: 28...Rbd8 29.Ke2 Rd4 30.b3 a4 31.h6 axb3 32.axb3
Re4+ 33.Kf3 Rxc4 34.bxc4 Rb8 35.Ke4, and White is in time to hold the enemy b-pawn. Victory is assured!
On the other hand, 26.Rxa6 may lead to unnecessary complications if Black plays 26...Ra8, meeting 27.Rf6 with
27...Bb1. True, White wins after 28.Nf5+ Kd7 29.Bd5, if we can believe our silicon friend: 29...Ra5 30.Bxf7 Bxa2
31.Be6+ Bxe6 32.Rxf8 Ra1+ 33.Kf2 c4 34.Nd4. Is it so, dear reader? Verify this on your own.
However, Torre preferred
26...Rbd8
and here, Polugaevsky missed a unique opportunity to create an everlasting masterpiece of the art of chess –
Selected Examples
In this chapter, you will encounter five treasures of the art of chess.
Three nominations:
A) the best attack of the twentieth century;
B) the best strategic attack in the post-Capablanca era;
C) the three best games of the computer chess age (irrational chess).
For (A), I quickly decided on the game Serper – Nikolaidis. Isn’t that because it was played in my hometown? Was I
being objective?
That’s not for me to judge. See for yourself – present any sort of counter-example. You, dear reader, have that
opportunity... Perhaps you will find something no less notable, maybe even in this very book – there are 125 annotated
examples, and 125 not commented by me. Go to it!
With regard to nominations for (B), I admit that I was totally non-objective: I like Karpov a lot! I consider him one of
the most elegant of the elegant strategic chessplayers. He’s a gifted, intuitive player...
From the games of the twelfth World Champion, I have selected a treasure forgotten by Karpov himself. Karpov forgot
to include his game with Miles in his collected games (see My Best Games (100 Victories in 30 Years), Moscow, Astrel
AST, 2002)! Of course, I hastened to correct this oversight.
For (C), without thinking too hard I limited my consideration to a very narrow circle of nominees: the trio of post-
Karpov world champions. My subjectivity is obvious here – so go ahead and protest! Find 10, 20, or 50 (who will find
more?) counter-examples. The more, the better; the gladder you will make me... Why?
Because then and only then will I know that you really – not in words, but in deeds – have absorbed my simple
arithmetic of the game of chess. And only then will I understand that you’re ready to storm the heights of modern
“higher mathematics” – the first step is behind you, and before you there lies... the chess abyss!
How does a strong modern grandmaster, armed to the teeth with chess arithmetic, beat another strong grandmaster, also
armed with the latest discoveries of chess knowledge?
Answer: by injecting into the chess system the “gene” of irrationality. By, Caissa permitting, moving stepwise from the
simple to the complex – that is, from order into chaos. And chaos is the element in which the goddess of the chess
universe resides...
Our age is the age of the predominance of chaos over order. We chessplayers hungrily colonize hitherto inaccessible
areas of the chess abyss – thanks to powerful computer programs! They are the locomotives of progress.
Are you ready?
16.a4
A mistake. 18...h3, ignoring White’s threat of invading on d6 with a knight, was stronger. The line runs 19.Ne4 hxg2
20.Nd6+ Qxd6 (20...Kd8!?) 21.cxd6 gxf1Q+ 22.Bxf1 Nd7 23.Rc1 0-0, with a complicated and approximately equal
position.
Another interesting variation after 18...h3 goes 19.d6 Qd8 20.c6 hxg2 21.Rfd1 Be6 22.Qxb4 Nd7 23.cxd7+ Bxd7
24.Kxg2. Here Black probably has full compensation for the pawn, as his king feels more comfortable.
Nor does White gain anything from 19.gxh3 or 19.g3.
19.d6 Qc6
The decisive error. Other bad moves are 19...Qd7, or 19...Qb7 20.c6+-; however, 19...Qa5 would have allowed him to
carry on the fight. For example, 20.Bc4 f4 21.Qd5 Ra7 22.Bd2 Be6 23.Qc6+ Bd7 24.Qb6 Qxb6 25.cxb6².
20.Bb5
23...f4 24.R1a7
Black is no better off after 24.R1a6 Qxa6 25.Rxa6 Bxa6 26.c6 Bc8 27.Bb6+-.
24...Nd7
One possible line (Serper) was: 24...fxe3 25.Qd5!! exf2+ 26.Kxf2 Qxd5 27.Rxc8#.
25.Rxc8+
A new and crushing wave of the attack. Serper is outstanding – this is his true element, and he is truly terrifying in it!
Fritz and Rybka unanimously prefer 28.Ne4, followed by 28...e2 29.Kf2 Nhf6 30.Ke1+-. Ingenious – but superhuman!
Serper’s variations are pretty: 29...Qa6+ 30.Kxf2 Qe2+! 31.Kxe2 Nf4+ 32.Kf1 Nxe6 33.c6 Kg8 34.Re7!+-; 29...Ng3+
30.hxg3 Qxd7 31.Qxd7 hxg3 32.Qe7+ Kg8 33.Qe8+ Kh7 (33...Bf8 34.Qxg6+ Bg7 35.Qxg3+-) 34.d7+-.
30.Rf7+
35.c6
35...e4 36.c7
36...e3 37.Qd5+ Kf6 38.Qd6+ Kf7 39.Qd5+ Kf6 40.Qd6+ Kf7 41.Qxe7+
The final blow. Previous to this, White had sacrificed a knight, bishop, rook, another bishop, another knight, and another
rook. In other words, in this game, White sacrificed the entire set of pieces!
41...Kxe7 42.c8Q Bh6 43.Qc5+ Ke8 44.Qb5+ Kd8 45.Qb6+ Kd7 46.Qxg6 e2+ 47.Kxf2 Be3+
48.Ke1
18.g3
A straightforward increase in ∆(move); but with this, White considerably increases both the third and the fourth factors
of his position.
He didn’t need this! Better was 19...Rxb7 20.Rxb7 Qd5 (… 21...c5) 21.c4 Qd6² (Karpov).
20.Rd7 Be3
Spectacular, but ineffective. Black is also worse after 20...Rxc3 21.Qxc3 Qxb1 22.Qxc6 … d5.
Clearly, Black proceeds in “Tal” style – attacking the target on c3. Does he have a right to do this?
You know the answer!
The parameters of the position are: m = 1, t = 48/37, and a “plus” in the third factor.
The diagnosis is unanimous – “Tal”!
23.d5
This move gives rise to primordial horror: it is pretty, and it’s in “Tal” style! It satisfies every requirement of the
position, and Miles deserves to be punished!
23...exd5
27.Qxg6+
Right! “Tal” has almost run out, and the “Capablanca” that replaces it welcomes exchanges.
27...hxg6
28.Ne5
Let’s examine this position from Black’s viewpoint: m > 1, t = 32/25 = 1.28, inferiority (by how much?) in the safety
factor, ∆k << 0, and ∆(28.Ne5) = ~-1.12.
Diagnosis (for Black): “Capablanca”? CP Algorithm?
Obviously, Black must play carefully.
28...Bg5
Loses. The bishop should have huddled together with the king: 28...Bg7 – a recommendation by the then-World
Champion. Here are some possible variations:
a) 29.Nxf7 Ra2 30.g4 d4 31.g5 d3 32.Kg3 Re2 (32...d2?? 33.Nh6+ Bxh6 34.gxh6+-) 33.Nh6+ Bxh6 34.Rh7= (Karpov);
b) 29.Nxg6 Bf6 30.Nxf8 Bxe7 31.Ne6² (Karpov); however, in the tempo endgame coming up, it looks like Black can
draw: 31...fxe6 32.Rxe7 c5 33.Rxe6 c4 34.Rc6 Rc3 35.h4 d4 36.Rc8+ Kh7 37.Rc7+ Kg8 38.g4 d3 39.Rd7 Rc2 40.h5 d2
41.g5 c3 42.h6 Rc1 43.Rd8+ Kh7 44.Rd7+ Kg8 45.Rd8+=;
c) 29.Nxc6!? Bf6 30.Re2 Rc8 31.Nb4² (Karpov).
In this last variation, White has good winning chances.
The most favorable scenario for White is: two rooks + knight + the f-, g-, and h-pawns against two rooks, bishop, and
the f- and g-pawns. I should add that the d-pawn is doomed, and we have good chances to win it while retaining the
minor pieces.
Now for the most favorable scenario for Black. Here it is: rook + the f- and g-pawns versus rook and three white pawns
on the kingside.
In the former case, White wins; while in the latter case, it’s a draw!
There is a third, intermediate, and the most probable scenario: a four-rook ending with f- and g-pawns for Black and f-,
g-, and h-pawns for White. The question is: is this a winning ending?
There is no answer: it’s a playable position!
After the game move, however, Black should lose.
29.Rxf7 Rxf7 30.Nxf7 Bf6
34.Ng5
34...Ra8
35.Rxc6 d4 36.Rxg6
The d-pawn will fall, too, because he cannot play 38...Re4 in view of 39.Rd8+ Kh7 40.Ng5+. The d-pawn is the fifth
consecutive one (!!) on the martyrs’ list of soldiers.
12.e4
Kramnik considers that after this move, Black “surprisingly loses” (V. Kramnik and Y. Damsky, Breakthrough). And
this categorical verdict of the fourteenth World Champion is endorsed by both Rybka and Fritz. Bravo!
12...fxe4
12...dxe4 loses to 13.Ng5, while after 12...Bxf3 13.Bxf3 fxe4 14.Nxe4 dxe4 15.Qxe6+ Qe7 (15...Be7 16.Bh5++-)
16.Rxe4 Kd8 17.Qb3 Black is defenseless, for example 17...Qf8 18.Bg5+ Nf6 19.Re6 (… 20.Rxd6+ Qxd6 21.Bf4+-)
19...b5 20.Rc1 Rb6 21.Bxh6 gxh6 22.Qc2 Bc5 23.Re3+-.
How did the game go?
13.Ng5
“I don’t remember the details anymore, but it seems that, during the game, first I looked at the ʻnormal’ 13.Nxe4, but
then found a more perfect continuation. It was just this that my opponent wasn’t expecting” (Kramnik).
13...Bf7
“On 13...Qxg5 14.Nxe4 Qe7, White has either 15.Bg5! Nf6 (after 15...Qf8, there follows 16.Qxd5! exd5 17.Nxd6#!)
16.Nxf6+ gxf6 17.Bxh6 and the attack now continues for free; or 15.Nxd6+ Qxd6 16.Bf4, picking up both the rook on
b8 and the pawns on d5, e6, and b7” (Kramnik).
After 13...Qe7 White also wins with either 14.Nxe6 (Kramnik) or 14.Ndxe4 (Rybka): 14...dxe4 15.Nxe6 Qf7 16.Bxe4
Rf8 17.Bf3 Qxf3 18.Ng5+.
Otherwise – let’s say, after 15...Qe7 or 15...Ng8 – White takes the e4-pawn with the rook and Black is defenseless.
Please verify this for yourself!
16.Qxe6+ Qe7
Black resigned, as the ending after 18...Nf6 19.Rxe7 Nxd5 20.Re6 offers no hope of salvation. 1-0
It is amazing that with such a starting value in the “t” parameter (~0.78), Black’s position was hopeless. The outcome
was decided by a pair of Tal-type blows. Why was this so?
The answer is obvious: these were all intrigues by the unpredictable third parameter! It’s definitely not good to take it
lightly...
A rhetorical question: can we ever restrain this strong-willed parameter?
No. 124: Kasparov – Shirov
Horgen 1994
17.Rxb7
The radical and, of course, not unquestionable solution. Kasparov sacrifices rook for bishop in order to clear the b4
square for the pawn. A “normal” treatment of this not-too-complicated position is also possible: 17.Nce3 … 18.Nc4,
when White is slightly better.
17...Nxb7 18.b4
There was a reason for the exchange sacrifice. His compensation (partial? full? excessive?) is the bad knight on b7.
The parameters of the position, from Black’s perspective, are: m > 1, t = 25/36 = ~0.69, and a “micro-minus” because of
the king (first pass) plus a “micro-minus” in the b7-knight (second pass), ∆k < 0, and ∆(18.b4) = ~-0.25.
The diagnosis is: the Petrosian Algorithm (the dynamic branch? the strategic branch?).
The standard question is: what should we do?
The no less standard answer: play according to the requirements of the Petrosian Algorithm.
Getting somewhat ahead of myself, I note that Shirov at first plays practically flawlessly. Let us see...
18...Bg5
This move I have no quarrel with – the bishop evades the white knight’s fire, which means that Black can now, at any
time he wants, busy himself with improving his queen’s position. The queen is obligated (consider it an order!) to leave
the d8 square in order to resolve the problem of the b7-knight – ...Nb7-d8-e6.
19.Na3
The knight is “elevated” to the c4 square – shooting for the target on d6.
19...0-0
Now the black king also feels safe. The further from a knight on d5, the more peacefully he sleeps.
20.Nc4
20...a5
The benefits of this move are more than evident: after the exchange on b4, the total mobility of the black pieces grows
considerably, as does the compactness of his position.
Its drawback is clear, too: the b-pawn becomes passed! In addition, after ...a6-a5 and ...a5xb4, White can stop thinking
about the a5 and c5 squares and push his pawn to b6. True, with this the black knight is liberated, but look at that pawn
on b6!
Then the question becomes: wouldn’t it have been better for Shirov to delay ...a6-a5 – maybe with 20...Ra7 … ...Qd8-
d7(b8), ...Nb7-d8?
I cannot answer that, because then Black risks dropping the a-pawn.
21.Bd3
Aimed at preventing a possible ...f7-f5. But 21.Be2 looks no worse – trading the e4-pawn for the f5-pawn is a double-
edged sword. And from e2 the bishop could go to g4. On the other hand, the white queen’s path to h5 would be blocked!
21...axb4 22.cxb4
The parameters of this position (for Black) are: m > 1, t = 35/42 = ~0.83, parity, or a “micro-minus” (the knight at b7) in
the third factor of the position, ∆k > 0, and ∆(22.cxb4) = ~-0.55.
Diagnosis: the CP Algorithm or even the left side of “Capablanca.” Let me add that Black’s bailiwick is on the left side,
not the right side, of the spectrum of all chess positions, and therefore he does not have the right even to think about
“Tal.” Here, for him, “Tal” would be a crime!
I consider my emotions here to be quite valid, considering that Shirov stepped over the invisible line...
22...Qb8
He needed to put the rook on a7 without delay and then play ...Kh8 (or maybe leave it on g8), followed by ...Qd8-
d7(e8), ...Bg5-h6, ...Nd8-e6, and ...g7-g6 with ...f7-f5 (or leaving these pawns alone). Black is compact, and he has lots
of choices for defense. He is relegated to playing a waiting game, but it seems to me that he would have good prospects
for survival!
Don’t torture yourself with such questions during the game! Play the position, strive within the limits of your own
strength not to offend the gods of chess... And then Caissa will smile upon you!
23.h4
To parry the straightforward 25.Nd7 with the threat of mate on f2 after 25...Qa7.
25.0-0
25...Rd2
A serious tactical, but also strategic, error – Black is not allowed to play according to “Tal”! He has to glue the queen to
the king at once with 25...Qe8 (26.Qb3 Ra7).
26.Qf3
Returning the favor! The “unexpected” 26.Qb1 (Rybka) 26...Qa7 27.Nc4 Ra2 28.Nc3 wins a rook for the knight. Alas,
spots appear on the sun too!
26...Qa7
No argument here.
27.Nd7 Nd8
But with this move, I have a thousand arguments. Shirov probably had a hallucination. The cool-headed Rybka tiptoes
along a narrow path to the draw: 27...Ra8 (fearlessly!) 28.Ne7+ Kh8 29.Qxf7 Rxd3 30.Nf8 (is that it?) 30...Qa2
(you’ve got to be kidding!):
Position after 30...Qa2 (analysis)
Here, perpetual check starting with 31.N(either)g6+ is probably the proper outcome for such a wild encounter...
After the text, however, there is no hope for Black. Kasparov concludes the game in outstanding fashion – on practically
every move, he plays Rybka’s top line!
Getting out of the pin. Among other things. White threatens 31.Qd7 with a crush.
32...Rxc4
Quite elegant.
9...Nb4
At that time – that is, at the dawn of the encroachment by powerful chess engines into the lives of professional
chessplayers – this move was called an opening novelty. A “Tal”-style move, almost trivial – a temporary (but is it?)
piece sacrifice.
11...g6
A spendid move that makes an indelible impression. Anand tears away the grubby sort of “Tal” – 11...Nf2, etc. –
preferring to expand the TC Algorithm to the TCP Algorithm!
11...Nf2 and 11...g6 – these are, correspondingly, the first and second lines from Rybka (13 ply) and Fritz (18 ply).
Check it out. And go further. Don’t be surprised to find surprises...
Chess is boundless, and the position after 11...g6 is boundless in its complexity!
12.b3
12...Bg7 13.bxc4 Nxf4 14.Nxf4 Bxe5 15.Nfe2 b4 16.Qa4+ Qxa4 17.Nxa4 Bxa1 18.Nxc5 0-0 19.Nd3 a5 20.g3 Bg7
21.Bg2 Ba6 22.c5 Rac8 23.c6 Rfd8 24.Rc1 Bh6+ 25.Nef4 Bxd3 26.Kxd3 e5 27.Kc4 exf4 28.Re1 fxg3 29.e5 Bf4
30.hxg3 Bxg3 31.Re3 Bf4 32.Re4 Bh2 33.Bh3 Rc7 34.Re2 Bg3 35.Re3 Bf4 36.Re4 g5 37.Kc5 Re7 38.Kd4 f6 39.d6
Bxe5+ 40.Rxe5 Rxd6+ 0-1
Show in Quiz Mode
Chapter 6
25.Bxf6+ 1-0
20...Rd1+ 0-1
34...Kg8 35.h4 Rd3 36.Rf5 Bd2 37.Rf3 Rxf3 38.gxf3 Kf7 39.Kg2 Ke6 40.Nd4+ Ke5 41.Nc2 h5 42.Kf2 Bb4 43.Ke2
Be7 44.Ne3 Bd8 45.Ng2 Kd4 46.Kd2 g5 47.hxg5 Bxg5+ 48.f4 Bd8 49.Ne3 h4 50.Ke2 h3 51.Kf3 h2 52.Nf5+ Kc3
53.Kg2 Kxb3 54.Nd4+ Kc4 0-1
35.Kh2 Bf7 36.Ne4 Qxa4 37.Nf6+ Kg7 38.Nd7 Rc8 39.Qf6+ Kg8 40.Re7 Qb3 41.Ne5 Kf8 42.Nxf7 1-0
40.h3 Rc6 41.Ra7 Kg8 42.Rhg7+ Kf8 43.Raf7+ Ke8 44.Ra7 Kf8 45.Rgf7+ Kg8 46.Rfb7 Rf4 47.Kh2 Rf2 48.h4 Rff6
49.Kh3 Rc3+ 50.g3 Rcc6 51.h5 Rf1 52.g4 Rf4 53.Ra8+ Rf8 54.Rxf8+ Kxf8 55.g5 a5 56.h6 Kg8 57.Kg4 Rc8 58.Kh5
1-0
35...h4 36.Rc7+ Kg6 37.gxh4 Nh5 38.Rd7 Nxf4 39.Bxf4 Rf3+ 40.Kg2 Rxf4 41.Rxd5 Rf3 42.Rd8 Rd3 43.d5 f4 44.d6
Rd2+ 45.Kf1 Kf7 46.h5 e3 47.h6 f3 0-1
27.b3 cxb3 28.Bxb3 Nf8 29.c4 Qd7 30.Qc2 Qb7 31.cxb5 axb5 32.Ng4 N6d7 33.Qd3 Qc6 34.Qe3 Kf7 35.Nxe5+
Nxe5 36.Qf4+ 1-0
21.Bxg6 hxg6 22.g3 Qf6 23.Rxe5 Qxe5 24.Rd1 Re8 25.Rd2 Qf6 26.Kg2 Re5 27.Nf3 Ne4 28.Re2 Rf5 29.Qd3 Kh7
30.Qe3 Qa6 31.Nd4 Rf6 32.f3 Nc5 33.a3 Ne6 34.Rd2 Nxd4 35.Qxd4 Re6 36.Kf2 Qb6 37.Qxb6 Rxb6 38.Ke3 Rd6
39.Kd4 b6 40.Re2 g5 41.b4 f6 42.Re7 a6 43.Ra7 b5 44.Kc5 Re6 45.Kxd5 Re3 46.Kd4 1-0
20.h4 Ng4 21.h5 Nxe5 22.dxe5 Qc8 23.Re3 Bg5 24.Rh3 Qc7 25.Re1 Rad8 26.Qc2 Re7 27.Re4 Red7 28.Rg3 Bh6
29.hxg6 hxg6 30.Rh4 Bg7 31.f4 Kf8 32.Kh2 Rd5 33.Rh7 R8d7 34.Bxg6 fxg6 35.Qxg6 Rf7 36.Qh5 Bxe5 37.Qh6+
Rg7 38.fxe5 Qxe5 39.Qxg7+ 1-0
61.Be8 Kd8 62.Bxg6 Nxg6 63.Kxb6 Kd7 64.Kxc5 Ne7 65.b4 axb4 66.cxb4 Nc8 67.a5 Nd6 68.b5 Ne4+ 69.Kb6
Kc8 70.Kc6 Kb8 71.b6 1-0
19...Qb6 20.Rab1 Qb4 21.Qf1 c4 22.Re2 b5 23.axb5 axb5 24.Kh1 Bxc3 25.bxc3 Qxc3 26.Rxb5 Qd3 27.Qe1 c3
28.Rb1 Nc5 0-1
23.Nxe6 Bxe6 24.Bxd5 Rd8 25.e4 Bg4 26.Rd3 Kh8 27.Qe5 Bf6 28.Qc7 Rd7 29.Qc6 Bh3 30.Be6 1-0
20...g5 21.fxg5 Ng6 22.Qg3 f4 23.Qg4 Nxg5 24.h4 Nf7 25.h5 Nfxe5 26.Qe6+ Kg7 27.Rh1 Qg5+ 28.Kf1 f3 29.Rg1
Qg2+ 30.Rxg2 fxg2+ 31.Kg1 Nf3+ 32.Kxg2 Nf4+ 33.Kxf3 Nxe6+ 0-1
16.Nc5 Be8 17.b4 Kg7 18.0-0 Rf8 19.Qa6 fxe4 20.Nd2 e3 21.Nde4 exf2+ 22.Rxf2 Bg5 23.Rxf8 Kxf8 24.Nxg5 hxg5
25.Qb7 1-0
26...f5 27.Rxb8+ Kf7 28.Bd4 Ne5+ 29.Bxe5 Qe4+ 30.Kg3 Qg4+ 31.Kh2 Rxh4#
19.Bg6 Bxe5 20.fxe5 fxg6 21.exf6 0-0 22.f7+ Kh8 23.Rf4 h5 24.Be7 Qc7 25.Qh4 Bb7 26.Re1 Qc6 27.Qg5 Kh7
28.h3 1-0
16.g3 Qf5+ 17.Kg2 Bd6 18.Qd1 g6 19.Bd3 Qe6 20.Rb1 Nc6 21.Rxb7 Rab8 22.Rxb8 Rxb8 23.Ne2 Kg7 24.Qa4 Ne7
25.Rb1 Rxb1 26.Bxb1 Bb8 27.Bc2 h5 28.Qb5 Bc7 29.h4 a6 30.Qb7 1-0
14...Bxc3 15.bxc3 c5 16.Qc2 Qe7 17.Rae1 Rfe8 18.Bh6 f6 19.f4 Qf7 20.Qd3 Re7 21.f5 g5 22.h4 g4 23.Qd5 Qxd5
24.cxd5 Rae8 25.Rf4 Ne5 26.Re2 Kh8 27.Rff2 Nf7 28.Bf4 Rxe4 29.Rb2 Rc4 30.Rb7 Re7 31.h5 h6 32.Rb3 a5 33.Kh2
a4 34.Rb8+ Kg7 35.Bd2 Ng5 36.Rf4 g3+ 37.Kxg3 Ne4+ 0-1
15...Qb3 16.Qe2 Ba4 17.Be3 Kb8 18.Rad1 Qc2 19.Rd2 Qf5 20.Rf1 g5 21.h3 h5 22.Nh2 Rdg8 23.g4 Qg6 24.Bf3
hxg4 25.Bxg4 Nc6 26.f3 Bd8 27.Bf2 Ne7 28.Re1 Rh6 29.Nf1 Rgh8 30.Bg3 Rxh3 31.Bxh3 Rxh3 32.Qg2 Qh7
33.Ne3 Ng6 34.Ng4 Nf4 35.Bxf4 gxf4 36.Kf1 Rg3 37.Qf2 Qh3+ 38.Ke2 Rg2 39.Rg1 Rxf2+ 40.Nxf2 Qh7 41.Rh1
Qg6 0-1
34.Re1+ Be5 35.d6+ Ke6 36.Qb3+ Kf5 37.Qd3+ Kg5 38.Qe3+ Kf5 39.Qe4+ Ke6 40.Qc4+ Kxd6 41.Rd1+ Ke7
42.Rxd7+ Kxd7 43.Qxa6 Rb8 44.Qa7+ Kc6 45.Qxh7 Rb2 46.Qxg6 1-0
33.b3 b5 34.Rxa6 Rd6 35.Qxf5+ Qg6 36.Qxg6+ Kxg6 37.Ra5 Rb6 38.g4 c6 39.Kg3 h5 40.Ra7 hxg4 41.hxg4 Kf6
42.f5 Ke5 43.Re7+ Kd6 44.f6 Rb8 45.g5 Rf8 46.Kf4 c4 47.bxc4 bxc4 48.Kf5 c3 49.Re3 1-0
18.Nb1 Qb7 19.Nc3 Nc7 20.Nb5 Qc6 21.Nxc7 Qxc7 22.Qb5 Ra8 23.c3 Rxa1 24.Rxa1 Rb8 25.Ra6 Bf8 26.Bf1 Kg7
27.Qa4 Rb7 28.Bb5 Nb8 29.Ra8 Bd6 30.Qd1 Nc6 31.Qd2 h5 32.Bh6+ Kh7 33.Bg5 Rb8 34.Rxb8 Nxb8 35.Bf6 Nc6
36.Qd5 Na7 37.Be8 Kg8 38.Bxf7+ Qxf7 39.Qxd6 1-0
24.Nh6+ gxh6 25.Bxh6 Rf7 26.Rd8 Ne7 27.Rc7 Ng6 28.Rcc8 e5 29.f4 Bd7 30.Ra8 Bh3 31.Kf2 b5 32.Rdb8 exf4
33.gxf4 Bd7 34.h4 Bc6 35.h5 Bxa8 36.hxg6 hxg6 37.Rxa8 f5 38.Kg3 a6 39.Kh4 Rg7 40.Kg5 1-0
17...Ng5 18.Nxg5 hxg5 19.dxe5 dxe5 20.Bxb6 axb6 21.Qe3 b5 22.Nf1 g4 23.hxg4 Bxg4 24.Rxd8 Rxd8 25.g3 Nf4
26.Nh2 Qg5 27.Kh1 Rd6 28.f3 Bh3 29.Rg1 Nd3 30.Qa7 Qd2 31.Bxd3 Rh6 32.g4 Bxg4 33.Rg2 Bxf3 34.Bf1 Qf4
35.Qg1 Bxe4 36.Be2 Rg6 37.Bf1 Qd2 38.b4 Qxa2 0-1
Position after26...f6
Show/Hide the answer
m = 1, t = ~1.46, “+” ... T
27.d7 Bc6 28.h4 Bxd7 29.h5 gxh5 30.e4 e5 31.f4 exf4 32.Rd6 Qe8 33.Bxf6 Rf7 34.Rd5 1-0
16.Nc4 Qe6 17.Rad1 cxb5 18.Qc7 Bd7 19.Nd6+ Ke7 20.Nf5+ gxf5 21.exf5 Rac8 22.Rxd7+ Qxd7 23.f6+ Nxf6
24.Re1+ Ne4 25.Rxe4+ Kf6 26.Qxd7 Rfd8 27.Qg4 1-0
17.exd4 Nd7 18.Qf4 Bxf3 19.Bxf3 Nxd4 20.Bxd4 cxd4 21.Qxd4 Nc5 22.Qxd8 Rxd8 23.Ra1 Kf8 24.Ra6 Nxa6
25.bxa6 Rd7 26.Kf1 Ra7 27.Bb7 Ke7 28.Ke2 Kd6 29.d4 Kc7 30.Kd3 Rxb7 31.axb7 Kxb7 32.Kc3 Kc6 33.Kb4 f6
34.g3 h5 35.f4 g5 36.fxg5 fxg5 37.h3 Kb7 38.Kb5 Kc7 39.g4 h4 40.c5 bxc5 41.dxc5 e5 42.Kb4 1-0
25.Qd1 b6 26.g3 Bf8 27.Bg2 Be7 28.Qh5 a6 29.h3 Qc6 30.Kh2 a5 31.f4 f6 32.Qd1 Qb5 33.g4 g5 34.Kh1 Qc6 35.f5
Bf7 36.e4 Kg7 37.exd5 Qc7 38.Re2 b5 39.Rxe7 Rxe7 40.d6 Qc4 41.b3 1-0
14...cxd4 15.Nxd4 Nxd4 16.Bxd4 Bc5 17.Qh4 Bxd4 18.Rxd4 f6 19.Bd3 h6 20.exf6 Rxf6 21.f5 Rcf8 22.Rg1 Nc5
23.fxe6 Nxe6 24.Rg4 Nf4 25.Qg3 Qe7 26.Rxf4 Rxf4 27.Ne2 Rf1 28.Nd4 Rxg1+ 29.Kxg1 Re8 30.h4 Qe1+ 31.Kh2
Qxg3+ 32.Kxg3 Kf7 33.Kf2 Kf6 34.g3 Bc8 35.c3 Bg4 36.Bc2 g5 37.hxg5+ hxg5 38.Bb3 Ke5 39.Bc2 Rf8+ 40.Kg2
Bd7 41.Nf3+ Kf6 42.Bb3 g4 43.Nd4 Ke5 44.Bc2 a5 45.Bd1 Ke4 0-1
43...b5 44.Bxc5 bxc4 45.dxc4 Bxc4 46.Qf2 Ng4 47.Qf4 Bxf1 48.Qxf1 Qxa2 49.Qf4 Qb1+ 50.Kg2 Qc2+ 51.Kh3
Qxc5 52.Kxg4 Qd4 53.Kf3 a5 54.Qc7 a4 55.Qa5 Qd3+ 56.Kf4 a3 57.h4 Qh3 58.Ke5 Qe6+ 59.Kd4 a2 60.h5 Qf6+
61.e5 Qf2+ 0-1
40...g5 41.hxg5 hxg4 42.fxg4 Kg6 43.Kg3 Kxg5 44.Bd2+ Kg6 45.Bc3 Be2 46.Nb2 Ne7 47.Na4 Nd5 48.Bb2 Bd3
49.Bd4 Bc2 50.Bb2 Kg7 51.Bd4 Kf8 52.Kf2 Ke7 53.Ke2 Kd7 54.Kd2 Kc6 55.Bb2 Nf4 56.Ke3 Nd3 57.Bd4 Kb5
58.Nb2 Nb4 59.Kd2 Nc6 60.Bc3 Kc5 61.Na4+ Kd5 62.Ke3 Bd1 63.g5 Bh5 64.Bb2 Bg6 65.Nb6+ Kc5 66.Nc8
Nb4 67.Nd6 Nd3 68.Bd4+ Kd5 69.Nb5 Kc4 70.Nd6+ Kb4 71.Nxf7 Ne1 72.Kd2 Nf3+ 0-1
Show in Quiz Mode
17.g4 Ne8 18.f5 Bf6 19.fxe6 fxe6 20.g5 Bxg5 21.Rxf8+ Kxf8 22.Ne4 Bh6 23.a5 bxa5 24.Nxc5 Qe7 25.Rf1+ Kg8
26.Qg4 Nc7 27.Ne4 Ne8 28.Bd4 Rc7 29.Bc5 Rxc5 30.Nxc5 Nc7 31.Ne4 Nc6 32.Nf6+ Kh8 33.Nd5 exd5 34.Qc8+
Ne8 35.cxd5 Nd8 36.Qxd8 1-0
16...exf6 17.Rac1 Qd6 18.Rc2 Rad8 19.Bf3 f5 20.Rd1 Rfe8 21.g3 Re7 22.Nc3 Rde8 23.Re2 Bf6 24.Rxe7 Rxe7 25.Na2
a5 26.Qb5 h5 27.b4 h4 28.Bg2 hxg3 29.hxg3 f4 30.g4 f3 31.Bxf3 Qf4 32.Qd3 Bxd4 33.Qxd4 Qxf3 34.Nc3 axb4
35.Qxb4 Bxd5 36.Nxd5 Qxd1+ 37.Kh2 Re6 38.Ne3 Qd6+ 39.Qxd6 Rxd6 40.Nc4 Rc6 41.Ne5 Rc3 42.Nd7 Rb3
43.g5 f5 44.Nf6+ Kf7 45.Nd5 Ke6 46.Nf4+ Kd6 47.Nxg6 Rb4 48.Kg3 Rg4+ 49.Kf3 Rxg5 50.Nf4 Rg1 51.Nd3 Ra1
52.Nb2 Kd5 53.Kf4 Kd4 0-1
20.g4 Na5 21.Qa4 Nc6 22.Bd2 Re6 23.Raf1 Rg6 24.g5 Re6 25.e4 g6 26.Qb3 Na5 27.Qc2 c4 28.exf5 gxf5 29.d4 Qe4
30.Qxe4 Rxe4 31.Rxf5 Rg4+ 32.Kh1 Rxf5 33.Rxf5 Nc6 34.Rc5 Re4 35.Rxc4 Kf7 36.Rc5 Kg6 37.Kg1 Re2 38.Bf4
Ne7 39.Kf1 Rb2 40.Ke1 Nf5 41.d5 Nh4 42.Kd1 Ng2 43.Bd2 Kf5 44.d6+ 1-0
18...Re3 19.h4 Qe7 20.Bxe3 fxe3 21.Qg2 b5 22.Bd3 Qe5 23.Be4 b4 24.Rc1 Rf4 25.Rc2 Ba4 26.Re1 bxc3 27.b3 Qd4+
28.Kc1 Bb5 29.Kb1 Rxe4 30.fxe4 Qd2 31.Re2 Qd1+ 32.Rc1 Bd3+ 33.Rec2 e2 0-1
13...g4 14.fxg4 fxg4 15.Qc2 Qe7 16.Qe4 Bb7 17.Qxg4 Qf7 18.Nh3 Rg8 19.Qe2 0-0-0 20.Qe3 Rg4 21.Be2 Re4
22.Qd3 Rg8 23.Rg1 Qf5 24.Be3 Ng3 25.Bf3 Rxe5 26.0-0-0 Bxf3 27.dxe5 Qxd3 28.Rxd3 Ne2+ 29.Kd2 Nxg1
30.Nxg1 Bg4 31.Rd4 Nb3+ 0-1
25.Na4 Qd8 26.Rg1 Nh6 27.Na1 Rf7 28.Nc2 exf4 29.exf4 Bxb2 30.Nxb2 Kf8 31.Na4 Ng8 32.Re1 Nf6 33.Rfe2 Rfe7
34.Rxe7 Rxe7 35.Rxe7 Kxe7 36.Ne1 Qc7 37.Nf3 Nd8 38.Qe2+ Kf8 39.Qb2 Ke7 40.Ng5 Nf7 41.Qe2+ Kf8 42.Qe6
1-0
33.Qb3 Re8 34.Ng4 Kg7 35.Kg2 f5 36.Nxh6 Kxh6 37.Bxf5 exf5 38.Qf7 1-0
49.a5 Nxf3+ 50.Kf2 Nxh2 51.c6 Ng4+ 52.Kf3 Nf6 53.Rb7 Nd5 54.Ke4 Nb4 55.c7 Rc8 56.Kd4 Kd7 57.Kc5 Nc6
58.a6 h5 59.a7 f5 60.a8Q Rxa8 61.c8Q+ Kxc8 62.Kxc6 1-0
15...Rxe3 16.fxe3 d6 17.Qd4 Nd7 18.Rf4 Qe7 19.Raf1 Ne5 20.Be4 Rf8 21.Bd3 Bc8 22.Qe4 g6 23.Rf6 Kg7 24.c4 Bd7
25.h4 Qxf6 26.Rxf6 Kxf6 27.Qd4 Re8 28.Kf2 Kg7 29.Ke1 f6 30.Kd2 Ng4 31.e4 Ne5 32.Qc3 Re7 33.Qa3 a5
34.Qb2 Nf7 35.Bc2 Re5 36.Qa3 Kf8 37.Ba4 Bg4 38.Qd3 Ke7 39.c5 dxc5 40.Qb5 Nd6 41.Qc6 Kd8 42.Bc2 Bc8
43.Qa8 g5 44.Qc6 gxh4 45.gxh4 Rh5 46.e5 fxe5 47.Qa4 e4 48.Qb3 Bb7 49.Qc3 Bxd5 50.Qf6+ Kd7 51.a4 Kc6
52.Qe7 Kb7 53.Kc3 Bc6 54.Kb2 c4 55.Kc3 Nc8 56.Qf6 Rc5 57.Qh6 Ne7 58.Qxh7 Nd5+ 59.Kd4 c3 60.Bxe4 Nb4
61.Bg6 Bxa4 62.h5 c2 63.Bxc2 Bxc2 64.Qf7 a4 65.h6 a3 0-1
17.Nxf7 Qxf7 18.fxe6 Qc7 19.Bf4 Bd6 20.Bxd6 Qxd6 21.Rf5 Qc7 22.a4 b4 23.a5 bxc3 24.axb6 Qxb6 25.Kh1 Rhe8
26.Qf1 Kb8 27.Rd1 Ka7 28.h3 Qb4 29.Ra1 Nxe4 30.Bxe4 Qxe4 31.Rf7 Qxd5 32.Rxa6+ Kb8 33.Qf4+ Rd6
34.Rxb7+ Kxb7 35.Rxd6 Qxb3 36.Qf7+ 1-0
26...Nd5 27.Rd1 Bg5 28.h4 Be3 29.Nd2 Bd4 30.Ne4 Qb6 31.Qc1 f5 32.Nc3 Qd6 33.Kg2 Qe5 34.Nxd5 Bxb2
35.Ne7+ Kf7 36.Nc6 Qd5+ 37.e4 Qxc6 38.Qxb2 fxe4 39.d4 e3+ 40.Kg1 Qf3 41.Rf1 Kg8 0-1
22.f3 Nc4 23.Nxc4 Bxc4 24.Qf2 Re8 25.e4 c6 26.Rd1 Rd7 27.Rxd7 Nxd7 28.Rd1 Qb7 29.Rd6 f6 30.f4 Re6 31.Rd2
Re7 32.Qd4 Nf8 33.Qd8 Rd7 34.Rxd7 Qxd7 35.Qxd7 Nxd7 36.e5 fxe5 37.Bxc6 Nf6 38.Bb7 exf4 39.gxf4 Nd5
40.Kf2 Nxf4 41.Ke3 g5 42.Bxa6 Kf7 43.a4 Ke7 44.Bxb5 Bxb5 45.axb5 Kd7 46.Ke4 Ne2 47.Bb6 g4 48.Bf2 Nc3+
49.Kf5 Nxb5 50.Kxg4 Ke6 51.Kg5 Kf7 52.Kf5 Ke7 53.Bc5+ 1-0
19...Ke7 20.Nd2 Ra8 21.Qb7 Qa5 22.Qb2 Rhb8 23.Qc2 Rxb1 24.Nxb1 Rb8 25.Nd2 Qa4 26.Rc1 Qxc2 27.Rxc2 Ba4
28.Nb3 Bc6 29.Nd2 Ba4 30.Nb3 h5 31.Ke1 h4 32.Kd2 Bc6 33.Kd3 f5 34.exf5 gxf5 35.Rf2 Ke6 36.Ke2 Ra8
37.Ke1 Ra4 38.g3 hxg3 39.hxg3 Be4 40.Rh2 Rxc4 41.Kd2 Ra4 42.Kc1 c4 43.Na1 Ra8 44.Rd2 Rg8 45.a4 Rxg3
46.Nc2 Bxc2 47.Rxc2 Rxe3 48.a5 Kd7 49.a6 Kc7 0-1
13.g3 cxd4 14.Bg2 Nf6 15.Qxd4 Rd8 16.Qf4 Bd7 17.Ne5 Be8 18.Rhd1 Qb6 19.b3 Rxd1 20.Rxd1 Rd8 21.Rxd8
Qxd8 22.Bxb7 Qa5 23.Qe3 Qxa2+ 24.Kf1 Qa5 25.Bf3 Qb6 26.c5 Bb5+ 27.Kg2 Qc7 28.c6 a6 29.Qc5+ Ke8 30.b4
Nd5 31.Bxd5 exd5 32.Qxd5 Kf8 33.f3 f6 34.Nd7+ Ke7 35.Qg8 Kd6 36.Nc5 Qe7 37.Qc8 Bxc6 38.Qxa6 Qe3
39.Qd3+ Qxd3 40.Nxd3 Be8 41.Kf2 Kd5 42.Nf4+ Kc4 43.Ne6 Kxb4 44.Nxg7 Bf7 45.Nf5 Kc5 46.Nh6 Bb3
47.Ng4 f5 48.Nf6 h6 49.Ke3 Kd6 50.Kf4 Ke6 51.Ng8 Bc2 52.Nxh6 Kf6 53.g4 fxg4 54.fxg4 Kg6 55.Nf5 Bd3 56.h4
Be2 57.h5+ Kh7 58.Kg5 Bd3 59.Kh4 Kg8 60.Nd6 Kg7 61.g5 Bc2 62.Nc4 1-0
22...c5 23.dxc5 Nxc5 24.Bxh6 Bxe5 25.Qxe5 f6 26.Qe3 Rd7 27.Bf4 e5 28.h6 g6 29.Bg3 Rfd8 30.h7 Rd2 31.Bxe5
Qe6 32.Bxf6+ Qxf6 33.Qc3 Qxc3 34.Rxc3 Rxf2 35.Rc2 Rdd2 0-1
14.Nc4 dxc4 15.Bxc4+ Nhf7 16.Rxd6 Qxd6 17.Nxe5 Be6 18.Rd1 Qe7 19.Rd7 Bxd7 20.Nxd7 Rfc8 21.Qc3 Rxc4
22.bxc4 Nd6 23.Qh8+ Kf7 24.Ne5+ Ke6 25.Qxa8 1-0
15.d5 Qxd5 16.Qxd5 exd5 17.axb5 Nd7 18.bxa6 Bc5 19.Be3 Bxe3 20.Rxe3 0-0 21.Ra5 c6 22.Nf5 Rfe8 23.a7 g6
24.Nd4 c5 25.Nb5 Rec8 26.e6 fxe6 27.Rxe6 d4 28.Re7 1-0
26...h6 27.Nc2 Be2 28.Re1 Nd3 29.Qxc5 dxc5 30.Rxe2 Rb1 31.fxg6 fxg6 32.Bf1 Rxc1 33.c4 Kg7 34.h4 h5 35.Rd2
Ne5 36.Kf2 Ng4+ 37.Ke2 Bc3 38.Rd1 Rxc2+ 39.Kd3 Rf2 40.Kxc3 Ne3 41.Rd7 Nxf1 42.Rxe7+ Kf6 43.Rc7 Nxg3
44.Rc6+ Kg7 45.e5 Re2 46.e6 Re3+ 47.Kd2 Re5 48.e7 Rxe7 49.Kd3 Re5 0-1
11.f3 Qe7 12.fxe4 Nxe4 13.Nxe4 Qxe4 14.Bd3 Qxg2 15.Bxf5+ Kxf5 16.Rf1+ Ke6 17.Qh5 Ne5 18.dxe5 Bxe5
19.Qf7+ Kd6 20.Rf6+ Kc5 21.b4+ Kxb4 22.Bd2+ Ka3 23.c5 d5 24.cxd6 Bg3+ 25.Rf2 Qh1+ 26.Ke2 Bg4+ 27.Kd3
Bd1 28.Bc1+ Ka4 29.Qc4+ Ka5 30.Bd2+ 1-0
32...g5 33.Nxg4 hxg4 34.Qf1 Rxh4+ 35.gxh4 g3+ 36.Rxg3 Qxf1 37.Rg1 Qf3 38.fxg5 Nf4 39.Bxf4 Qxf4+ 40.Kh3 Bf3
41.Rf2 Bg4+ 42.Kg2 Qe4+ 43.Kh2 Qe3 44.Rgg2 d3 45.g6 Qxe5+ 46.Rg3 Qd4 47.Rd2 Qf6 48.Rgxd3 Qxh4+ 49.Kg1
Qg5 50.Rg2 Qxg6 51.Kf2 Qf6 52.Rd5 Qh4+ 53.Kg1 f4 54.Rgd2 Qe1+ 55.Kh2 Bf3 0-1
31.Bf1 Rxc2 32.Rxe2 Rxe2 33.g6 hxg6 34.Qxg6+ Kd7 35.Qh5 Rxb2+ 36.Bxb2 Qb7 37.Bc4 Bc3 38.Qh7+ Kc8
39.Qxb7+ Kxb7 40.Bxd5+ 1-0
21.Kh1 Rc6 22.Be2 Qb7 23.Bf1 Rfc8 24.Qg4 Nd7 25.Bc4 Nf6 26.Qf3 h6 27.Bf2 Rxc4 28.bxc4 Rxc4 29.Bh4 Rd4
30.Rad1 Nc6 31.Bxf6 Bxf6 32.Nd5 Bg5 33.Rxd4 Nxd4 34.Qa3 Qc6 35.Qc3 Qb5 36.b4 Qe2 37.Qd3 Qa2 38.Rf1 h5
39.g3 f6 40.Kg1 h4 41.Rf2 Qa1+ 42.Kg2 Kh7 43.gxh4 Bf4 44.Kh3 Kh8 45.Ne7 1-0
27.Rxg7 Rxg7 28.Bxg7+ Kxg7 29.Qg5+ Kh8 30.Qe7 Qe4 31.Rxd7 Qe1+ 32.Kg2 Qe4+ 33.f3 Qc2+ 34.Kh3 Qf5+
35.Kh4 Kg7 36.Rd4 Rc8 37.Rf4 Qxe5 38.Rxf7+ Kg8 39.f4 1-0
16.b4 axb4 17.axb4 Ncxb4 18.Nfxd4 exd4 19.Ba3 Nxc2 20.Bxe7 Qxe7 21.Rxe7 Nxa1 22.Qa5 Nb3 23.Qd5+ Kh8
24.Qf7 Rg8 25.Re8 Be6 26.Rxg8+ Rxg8 27.Qxe6 Nac5 28.Qb6 h6 29.Bh3 Re8 30.Bf5 Kg8 31.Bg6 Rc8 32.Kg2 Kh8
33.Bf5 Re8 34.Bg4 Re5 35.Qd8+ Kh7 36.f4 f5 37.fxe5 fxg4 38.Qd6 Nxd3 39.e6 Ne1+ 40.Kf1 Nf3 41.e7 Nbd2+
42.Ke2 Ne4 43.Qf4 1-0
17.d5 cxd5 18.Nd4 Kf7 19.Ne6 Rhc8 20.Qg4 g6 21.Ng5+ Ke8 22.Rxe7+ Kf8 23.Rf7+ Kg8 24.Rg7+ Kh8 25.Rxh7+
1-0
20.b4 gxf4 21.gxf4 Nh4 22.bxc5 bxc5 23.Rb1 Rg8 24.Kf1 Ke8 25.Bh7 Rf8 26.Rb7 a6 27.Qb2 Rc8 28.Kf2 Ra8
29.Rg1 Qc8 30.Bc2 1-0
25.fxg6 f6 26.Qg5 Qd7 27.Kg1 Bg7 28.Rxf6 Rg4 29.gxh7+ Kh8 30.Bxg7+ Qxg7 31.Qxg4 1-0
22.Bc2 Bc8 23.Ba4 g5 24.Qg3 Nh5 25.Qf3 g4 26.Qd1 Re7 27.h3 Nf6 28.hxg4 Bxg4 29.f3 Bc8 30.Re3 Nh5 31.Qe1
f6 32.Qh4 Nf4 33.Ne4 Rf7 34.g3 Ng6 35.Qh5 Bb8 36.Bc2 f5 37.f4 Re8 38.Nf2 Rf6 39.Rae1 e4 40.Bxe4 fxe4
41.Nxe4 1-0
19.g5 Nxe4 20.f6+ Kf8 21.Bd3 Rd4 22.Rde1 Nd6 23.Qh5 e4 24.g6 fxg6 25.Qxh7 gxf6 26.Rhg1 g5 27.Rxg5 fxg5
28.Rf1+ Ke8 29.Qg7 Kd8 30.Rf8+ Ne8 31.Qxb7 Rc8 32.Qf7 1-0
11...f6 12.Bd3 fxe5 13.fxe5 Ndxe5 14.dxe5 Nxe5 15.Be2 Bd7 16.Nxe5 Qf2+ 17.Kd2 Rac8 18.Qb3 Bg5+ 19.Kd3 Rf4
20.Nf3 Be8 0-1
33...Bc8 34.a5 h6 35.Kf1 Kf7 36.Ke2 Ke8 37.Bc7 Kd7 38.Bb6 Ke7 39.Nd2 Nc3+ 40.Kf3 g5 41.Ne4 Bd7 42.Nxc3
Bxc3 43.Ke2 Bf6 44.Na3 Ba4 45.Nb1 Bc2 46.Nd2 Bc3 47.Nf3 Bb4 48.Bc7 Kd7 49.Bb6 Bb3 50.Kd3 Ke8 51.Bc7
Ke7 52.Bb6 Kd7 53.Ng1 Be1 54.f3 Ba2 55.Ke2 Bb4 56.Kd3 Bb1+ 57.Ke2 Bc3 58.h4 gxh4 59.Nh3 Be5 60.Nf2
Bg6 0-1
12.Qd2 Qxb2 13.Rb1 Qa3 14.Nb5 Qxa2 15.Nd6+ Kf8 16.Rd1 Qb2 17.Be2 Qb6 18.c4 d4 19.Bf3 a5 20.0-0 d3+
21.Kh1 Qd4 22.Nb5 Qc5 23.Qxd3 g6 24.Nd6 Nb6 25.Rb1 Kg7 26.Rb5 Qc7 27.Qd4 Nd7 28.f5 gxf5 29.Nxf5+ exf5
30.e6+ Ne5 31.Rxe5 f6 32.Rxf5 Rf8 33.Bd5 Qe7 34.Rh5 Kh8 35.Be4 Bxe6 36.Rxh7+ Qxh7 37.Bxh7 Kxh7 38.Qe4+
1-0
16.Rxc7 Bc6 17.R1xc6 bxc6 18.Rxf7 h6 19.Rb7 Qc8 20.Qc4+ Kh8 21.Nh4 Qxb7 22.Ng6+ Kh7 23.Be4 Bd6
24.Nxe5+ g6 25.Bxg6+ Kg7 26.Bxh6+ 1-0
13.Qb3 Qxa1 14.Bb2 Qb1 15.Nf3 Qxh1 16.Ne5 e6 17.Bxd7+ Rxd7 18.Qb8+ Rd8 19.Qb5+ Ke7 20.Qb7+ Kf6
21.Qxf7+ Kg5 22.Nf3+ 1-0
17.h3 Rfc8 18.g4 g6 19.gxf5 gxf5 20.Ne5 Ne8 21.Rg3+ Kh8 22.Kh2 Nf6 23.Rg1 Rc7 24.Bf3 Bc6 25.Qb3 Rg8
26.Bh5 Qf8 27.Rxg8+ Nxg8 28.Qg3 Bb5 29.Qh4 Nf6 30.Bf7 1-0
22.Kh4 Qxe5 23.Be8+ Kh8 24.Qxf3 Qf5 25.Qxf5 exf5 26.Re1 Kg8 27.Kg5 Nd8 28.Kf6 a5 29.Bf7+ 1-0
21...c6 22.Bf3 Kc7 23.Kf1 Re6 24.g3 Ng6 25.Bh5 Nxe5 26.f4 gxf4 27.gxf4 c5 28.Rd2 Nf3 29.Bxf3 Bxf3 30.Rxe6
dxe6 31.Rxd8 Kxd8 32.Na4 Kc7 33.Kf2 Bc6 34.Nc3 Kd6 35.Ke3 e5 36.Ne2 Bd7 37.a3 Be6 38.Ng3 a5 39.a4 Bd7
40.Ne2 Bc6 41.Nc3 Ke6 42.Ne2 Be4 43.Nc1 exf4+ 0-1
11.d5 exd5 12.cxd5 Nxd5 13.Rxd5 Bg7 14.h4 g4 15.Ne5 Nxe5 16.Bxe5 0-0 17.Bxg7 Kxg7 18.Qg3 f5 19.e4 Bb7
20.Rd6 Qe7 21.Bb5 Rad8 22.Rhd1 Bc6 23.exf5 Bxb5 24.Qxg4+ 1-0
14.Bd3 h6 15.Bh7+ Kh8 16.Rxg7 Kxg7 17.Qg6+ Kh8 18.Qxh6 Qe7 19.Bg6+ Kg8 20.Ke2 Qg7 21.Qh5 Rfd8 22.Rg1
Kf8 23.Rg3 Ke7 24.Bf5 Qf7 25.Bxf6+ Qxf6 26.Qh7+ Qf7 27.Rg7 Be8 28.Bg6 Qxg7 29.Qxg7+ Ke6 30.Bh7 Ne7
31.f4 Bb5+ 32.d3 Rf8 33.Qe5+ Kd7 34.f5 Rae8 35.f6 Nc8 36.Qxd5+ Nd6 37.Bg6 Re6 38.Bf5 Rfxf6 39.Bxe6+ Rxe6
40.h4 Ke7 41.Qc5 a6 42.Qxc7+ Kf6 43.Qd8+ Kf7 44.Kd2 Nf5 45.Qc7+ Re7 46.Qf4 Ke6 47.e4 Nd4 48.Qh6+ Kd7
49.Qf6 1-0
36.g5 hxg5 37.hxg5 Nxg5 38.Bc6 Rc8 39.Rh1 Kg8 40.Raa1 Nce6 41.Rh4 f6 42.Rg1 Kf7 43.Bd5 Ke8 44.Rh8+ Nf8
45.f4 Nge6 46.Kf3 f5 47.Rg6 Nc5 48.Bxg7 Rf7 49.Bxf8 1-0
25.Nh2 Bxf5 26.exf5 Nxh4+ 27.Kh3 Bxg5 28.Bxg5 Qxg5 29.Rg1 Qf6 30.Qxh5 Nxd5 31.Qxh4 Nf4+ 32.Kg4 Qxh4+
33.Kxh4 f6 34.c4 Kf7 35.Kg4 Rab8 36.Rgb1 Rh8 37.Nf3 g6 38.Nh4 gxf5+ 39.Nxf5 Ke6 40.Rxa5 Rbg8+ 41.Kf3
Rg5 42.Ng3 f5 43.Ra6 1-0
Show in Quiz Mode
17.Ne1 g5 18.g3 Nf5 19.Ng2 Qg7 20.f4 Nd6 21.Qe1 b5 22.fxg5 Rxf1+ 23.Bxf1 hxg5 24.Ne3 Rf8 25.Bg2 a5 26.Bd2
Qg6 27.Ng4 Rf5 28.Qe3 Kg7 29.Rb1 Kh7 30.a4 bxa4 31.Bc1 Rf7 32.Ba3 Qxc2 33.Rc1 Qf5 34.Bh3 Ne4 35.Ne5
Qf2+ 36.Qxf2 Rxf2 37.Nxd7 Ra2 38.Bc5 Nd2 39.Bg2 a3 40.Nf8+ Kh6 41.Re1 e5 42.dxe5 g4 43.e6 Nf3+ 44.Bxf3
gxf3 45.Kf1 1-0
19...b4 20.axb4 Bxe4 21.fxe4 Qxc2+ 22.Ka1 a3 23.Rb1 Qxe4 24.Bd2 Nd5 25.Nc1 Qxe1 26.Rxe1 Bf6 27.Rd1 Rc2
28.Nd3 Nb6 29.Bc1 Nc4 30.b5 Na5 0-1
24.Nb1 Qb7 25.Kh2 Kg7 26.c3 Na6 27.Re2 Rf8 28.Nd2 Bd8 29.Nf3 f6 30.Rd2 Be7 31.Qe6 Rad8 32.Rxd8 Bxd8
33.Rd1 Nb8 34.Bc5 Rh8 35.Rxd8 1-0
18...Kh8 19.Rc2 Rg8 20.Rcd2 g5 21.Bd4 Rg6 22.Nc1 Rcg8 23.Nd3 Qf8 24.Re1 g4 25.fxg4 e5 26.Be3 Nxg4 27.Nd5
Bd8 28.Nf2 Bh4 29.Ree2 Nxe3 30.Nxe3 Bxf2 31.Qxf2 Bxe4 32.Nf5 Nc5 33.Ng3 Ba8 34.Rd1 Ne6 35.Qxb6 Nf4
36.Rf2 Qh6 37.Kg1 Qh4 38.Qb3 Rh6 39.Rxf4 exf4 40.Qc3+ f6 41.Nf5 Rxg2+ 42.Bxg2 Qxh2+ 0-1
9.Be3 b4 10.Bxc5 Qxe2+ 11.Nxe2 Na6 12.Bxd6 Bxd6 13.Nd2 Rb8 14.Nc4 Be7 15.d6 Bf6 16.Nf4 Nc5 17.0-0 Ba6
18.Nd5 0-0 19.Nxf6+ gxf6 20.Bd5 Na4 21.Rfc1 Rfe8 22.Ne3 Rb6 23.Rc7 Rxd6 24.Rd1 Kh8 25.Nf5 Rb6 26.Bxf7
Nxb2 27.Rdxd7 Re1+ 28.Kg2 Bf1+ 29.Kf3 Be2+ 30.Kf4 Rb8 31.Bc4 1-0
17...Rd8 18.Rd1 d4 19.Bxb7 Qxb7 20.Ne4 Bg7 21.0-0 Qe7 22.Rfe1 Re6 23.Qd2 Rxe5 24.f4 Re6 25.f5 Bh6 26.f6
Be3+ 27.Rxe3 dxe3 28.fxe7 Rxd2 29.Nxd2 Rd6 30.Kf1 exd2 31.Ke2 b4 32.Rxd2 Rxd2+ 33.Kxd2 Kxe7 34.Ke3 a5
35.a4 f5 36.Kd2 f4 0-1
22.Qe1 Bxh4 23.Qa5 Be7 24.Qc7+ Ka8 25.Qa5+ Kb8 26.Qc7+ Ka8 27.Rfe1 Bd6 28.Qb6 Bb8 29.a5 Rd7 30.Re8
Qh2+ 31.Kf1 Qxg2+ 32.Kxg2 d4+ 33.Qxb7+ Rxb7 34.Rxh8 Rxb5 35.a6 Ka7 36.Rf8 Rxb2 37.Rxf7+ Ka8 38.a7 c3
39.Rf8 1-0
13.Bxe7 Nxe7 14.Nxd6+ Kf8 15.Ng5 b6 16.Qb4 h6 17.Qb3 Nd5 18.Ngf7 Be6 19.Nxh8 Bxh8 20.c4 Ne7 21.g4 Nd7
22.Qa3 c5 23.d5 Nxd5 24.cxd5 Bxd5 25.Rbd1 1-0
11.Bd5 Kh8 12.Ne4 Qd8 13.h4 f5 14.Ng5 Bb7 15.g4 e4 16.Ne2 Bxb2 17.Nf4 Qf6 18.gxf5 Bxa1 19.Nxg6+ Kg7
20.Nxe4 Bc3+ 21.Kf1 Qxf5 22.Nf4 Kh8 23.Nxc3 Rae8 24.Nce2 Rg8 25.h5 Rg5 26.Ng3 Rxg3 27.fxg3 Rxe3 28.Kf2
Re8 29.Re1 Rxe1 30.Qxe1 Kg7 31.Qe8 Qc2+ 32.Kg1 Qd1+ 33.Kh2 Qc2+ 34.Ng2 Qf5 35.Qg8+ Kf6 36.Qh8+ Kg5
37.Qg7+ 1-0
17.g6 hxg6 18.Rxg6 e5 19.Rxg7 exd4 20.Bxd4 Nc4 21.Rdg1 Nf6 22.e5 dxe5 23.fxe5 Nd7 24.e6 Qf4+ 25.Kb1 Qxd4
26.exd7+ Kxd7 27.Bf5+ Kc7 28.Qxe7+ Kb8 29.Bxc8 Rxc8 30.Qxf7 Nd6 31.Qd7 Qc5 32.Rd1 b4 33.Rxd6 1-0
20.e5 Ngxe5 21.Nc5 Qxd6 22.f4 Nd7 23.fxg5 Nxc5 24.gxf6 e5 25.Qh5 Rf7 26.Bh6 Ne6 27.Be4 Nf8 28.Qg5+ Kh8
29.Qg7+ 1-0
17.Rf1 f5 18.Rf3 Qxa2 19.Rfb3 fxe4 20.Qc3 Bd8 21.Qxc6+ Ke7 22.Qxa8 Qxc2 23.Qa7+ Bd7 24.Rb7 Ba5+ 25.Kf1
Rd8 26.Rd1 Bd2 27.Rc7 Rf8+ 28.Kg1 Be3+ 29.Qxe3 Qxc7 30.Qg5+ 1-0
15.Nd6 Nf8 16.Nxf7 exf1Q+ 17.Rxf1 Bf5 18.Qxf5 Qd7 19.Qf4 Bf6 20.N3e5 Qe7 21.Bb3 Bxe5 22.Nxe5+ Kh7
23.Qe4+ 1-0
23...Rc6 24.Ng4 Qd8 25.Rh3 Qf8 26.a4 Na3 27.Bxa3 Bxa3 28.Rg3 h5 29.Ne3 Bb2 30.Ra2 Bc1 31.Qg5 Qb4 32.Nf4
Bxe3 33.fxe3 Rc1 34.Nxg6 Rxd1+ 35.Kh2 fxg6 36.Qxg6 Qe7 37.Rf2 Qxh4+ 38.Rh3 Qxf2 0-1
11...c5 12.Bf3 cxd4 13.cxd4 Nd7 14.Bxa8 Qxa8 15.0-0 c3 16.Nf3 Rc8 17.Qa4 Nf6 18.Rfc1 Nd5 19.Qc2 a5 20.a3
Qa6 21.Bg3 Bh6 22.Re1 b5 23.e4 b4 24.Ne5 Bd2 25.Rf1 Qb5 26.a4 Qb7 27.Nf3 Bh6 28.h4 Nb6 29.d5 Bd7 30.Ra1
Nxa4 31.Nd4 Qb6 32.Nb3 Bb5 33.Rfb1 Nc5 34.Nxc5 Rxc5 35.Be5 Rc8 36.Rd1 a4 37.d6 b3 38.d7 Bxd7 39.Qd3
Bb5 40.Qh3 Qe6 0-1
25.Bxh5 Bxh5 26.Rxh5 Ng3+ 27.Kg2 Nxh5 28.Rxb7 a5 29.Rb3 Bg7 30.Rh3 Ng3 31.Kf3 Ra6 32.Kxf4 Ne2+ 33.Kf5
Nc3 34.a3 Na4 35.Be3 1-0
18.Rb1 Bc8 19.Ng5 Bf6 20.Qh5 Bxg5 21.Bxg5 Re8 22.Bf6 gxf6 23.Qh6 Nac5 24.Rbd1 Qb7 25.Rd4 Ne4 26.Rexe4
dxe4 27.Qxf6 Qc7 28.Nh6+ Kf8 29.Qh8+ Ke7 30.Nf5#
21.Ncd1 Bd7 22.Qa6 Qe8 23.Ba5 Ng6 24.Bc7 Bc8 25.Qa5 g4 26.fxg4 hxg4 27.hxg4 Nh4 28.Rb8 Qg6 29.Rxa8
Nxg4 30.Nxg4 Bxg4 31.Rxf8+ Kh7 32.Bxd6 Bh5 33.g4 Bxg4 34.Rh8+ Kxh8 35.Qd8+ Kh7 36.Qxh4+ Bh5+
37.Bg4 1-0
17.f5 e5 18.Be3 b4 19.axb4 Qxb4 20.Bg5 Qxb2 21.Rd3 Qxc2 22.Bd1 Qb2 23.Bxf6 Bxf6 24.Rxd6 Bb5 25.Rf2 Qc1
26.Rxf6 Rac8 27.Rc2 Qa1 28.fxg6 hxg6 29.Rd6 Rc7 30.Qxe5 Rfc8 31.Qd5 Kg7 32.Qd4+ Kh7 33.Nxb5 1-0
17.Rh5 Be8 18.Rxd8+ Kxd8 19.Qd2+ Bd6 20.Nd3 Qc7 21.g4 Kc8 22.g5 Bf8 23.Rh4 Kb8 24.a4 Be7 25.a5 Nd5
26.Kb1 Bd8 27.a6 Qa5 28.Qe2 Nb6 29.axb7 Bxg5 30.Nxg5 Qxg5 31.Rh5 Qf6 32.Ra5 Bc6 33.Nc5 Bxb7 34.Nxb7
Kxb7 35.Qa6+ Kc6 36.Ba4+ Kd6 37.Qd3+ Nd5 38.Qg3+ Qe5 39.Qa3+ Kc7 40.Qc5+ Kd8 41.Rxa7 1-0
9.dxc5 bxc5 10.0-0-0 Ne4 11.Qd3 Nxf2 12.Bxd8 Nxd3+ 13.Bxd3 Rxd8 14.Be4 d5 15.cxd5 Bb7 16.Ne2 Nd7
17.dxe6 Bxe4 18.e7 Re8 19.Rxd7 f6 20.Ng3 Bg6 21.h4 h6 22.h5 Bh7 23.Rh4 Kf7 24.Rc4 Rxe7 25.Rxe7+ Kxe7
26.Rxc5 Kd6 27.b4 Re8 28.Kd2 Re5 29.Rxe5 Kxe5 30.a4 Bg8 31.b5 Bb3 32.a5 Bc4 33.b6 axb6 34.axb6 Bd5 35.e4
Bb7 36.Ke3 Bc6 37.Nf5 Bxe4 38.g4 1-0
30.b4+ Ka4 31.Bc6+ Kb3 32.Qg3+ Kb2 33.Rb1+ Bxb1 34.Rxb1+ Kxb1 35.Qb3+ Ka1 36.Kc1 1-0
27.c3 a5 28.Nb5 g5 29.c4 Nf4 30.Rxd8+ Bxd8 31.Bxf4 gxf4 32.Qd3 Qb6 33.Nd6+ 1-0
26.Rfe1 axb3 27.Bxd5 cxd5 28.Re7 Be6 29.Rxe6 b2 30.Rae1 c2 31.Re7 Qxe7 32.Rxe7 1-0
23...Ba5 24.Rc8 Qb5 25.Rfc1 Qxb7 26.Qe8 Qxc8 27.Qxc8 Bd2 28.h3 h6 29.Qc4 Bxc1 30.Qxc1 Rxf2 31.Qc7 a6
32.Qa7 Rf6 33.a4 Rd8 34.a5 Rd1+ 35.Kh2 Rd2 36.Qb8+ Kh7 37.Qb4 Rff2 38.Qe4+ f5 0-1
18.Qd6 Ne8 19.Qxf8+ Kxf8 20.fxg4 f6 21.Ne4 Kg7 22.Bd1 Be6 23.Be1 Qh6 24.Bc3 Qg6 25.Ng3 Qd3 26.Kh2 Rd8
27.Be2 Qg6 28.e4 Kg8 29.Rf2 Nd6 30.d5 Qh6+ 31.Kg1 Qh4 32.dxe6 Qxg3 33.Rd1 Kf8 34.Rd3 1-0
13...g5 14.f5 Ne5 15.Qh3 Kd7 16.Be2 h5 17.fxe6+ fxe6 18.gxh5 g4 19.Qg2 Rxc3 20.bxc3 Bxe4 21.Qf2 Nxh5 22.Bf4
g3 23.hxg3 Nxf4 24.Qxf4 Rh1+ 25.Kf2 Rh2+ 26.Ke3 Bg5 27.Kxe4 Bxf4 28.gxf4 Rh3 29.Nxe6 Kxe6 0-1
27.h3 d4 28.Kh2 d3 29.Rc1 Qxf2 30.Rf1 Qd4 31.Rf5 e3 32.Rxa7+ Qxa7 33.Ra5 1-0
19.d4 f4 20.dxe5 g5 21.exd6 gxh4 22.e5 Nc6 23.b4 c4 24.a4 bxa4 25.Bxa4 Qb7 26.Bxc6 Qxc6 27.Nd4 Qd7 28.Qf3
Kh8 29.Ra5 Qf7 30.Rc5 Ra7 31.Qe4 Re8 32.Kf1 Rd7 33.Re2 Kg8 34.Ra2 Qh5 35.f3 Qg5 36.Rxa6 Bxh3 37.gxh3
Bxe5 38.Nf5 Qf6 39.Raa5 1-0
20...Qb6 21.Nd6 Nd7 22.Qg5+ Nf6 23.Rd3 Qb4 24.Qxe5+ Be6 25.Rhd1 Bh6+ 26.Kb1 Qf4 27.Qc5 f1Q 28.Nf5+
Kxf7 29.Qe7+ Kg6 30.Nh4+ Kh5 31.Qc5+ Bg5 0-1
15...Nh5 16.g4 Rxb2 17.gxh5 g5 18.Rg1 g4 19.0-0-0 Rxa2 20.Nef4 exf4 21.Nxf4 Rxf4 22.Qxf4 c3 23.Bc4 Ra3
24.fxg4 Nb4 25.Kb1 Be6 26.Bxe6 Nd3 27.Qf7 Qb8+ 28.Bb3 Rxb3+ 29.Kc2 Nb4+ 30.Kxb3 Nd5+ 31.Kc2 Qb2+
32.Kd3 Qb5+ 0-1
9.e5 Ng4 10.0-0 Nxf2 11.b4 Nh3+ 12.Kh1 Nf2+ 13.Rxf2 Bxf2 14.Nbd2 Ba7 15.Ne4 0-0 16.Nf6+ gxf6 17.exf6 Nd7
18.Qd2 Rd8 19.Qg5+ Kf8 20.Bf4 Qc6 21.Rd1 Nxf6 22.Rxd8+ Ne8 23.Ne5 1-0
16.Qc2 dxe4 17.Nxe4 Nxe4 18.Rxe4 g6 19.Bg5 Qd6 20.dxe5 Qe6 21.Nd4 Qd5 22.Bf6 c5 23.c4 Nxc4 24.Nf3 Na5
25.Rh4 Bg7 26.Be4 Qd7 27.Bxg7 Bxe4 28.Qc1 Bxf3 29.Qh6 f5 30.Qxh7+ Kf7 31.gxf3 Rg8 32.Bf8+ 1-0
16.d5 exd5 17.Qf5 dxe4 18.0-0-0 Rc7 19.Bg4 Bb5 20.Qxe4+ Kd8 21.Bxd7 Bxd7 22.Rhe1 Bh6 23.Qa8+ Rc8
24.Rxd7+ Kxd7 25.Qd5+ 1-0
13...axb5 14.Qxa8 Bd4 15.Nxd4 cxd4 16.Qxb8 0-0 17.Ke1 Qh4+ 18.g3 Qf6 19.Bf4 g5 20.c5 exf4 21.Qd6 Qg7
22.Bd3 Ne5 23.Kd2 f3 24.Bxb5 g4 25.Qe7 Ng6 26.Qg5 h6 27.Qh5 d3 28.Bxd3 Re8 29.h3 Re5 30.hxg4 Rxh5
31.gxh5 Ne5 32.Rae1 Qg5+ 33.Kc2 f2 34.Rd1 Qe3 0-1
17.Nxg7 Kxg7 18.Qh5 Rg8 19.Bxe6 bxc3 20.Bxf7 Qa5 21.Rh3 Kf8 22.Rf3 Ne5 23.Bd5+ Kg7 24.Rf7+ 1-0
12.Bh6 Bxh6 13.Qxh6 b4 14.Nb1 Qxa2 15.Ng3 Nb6 16.c5 Nc4 17.Rd2 Nxd2 18.Nxd2 Qa1+ 19.Nb1 Be6 20.cxd6
Qa5 21.Qg5 Qd8 22.dxe5 Nd7 23.Qh6 g5 24.Nf5 Bxf5 25.exf5 f6 26.Bc4+ Kh8 27.e6 Ne5 28.Rd1 Nxc4 29.e7 Qe8
30.Qxf8+ Qxf8 31.exf8Q+ Rxf8 32.Rd4 Ne5 33.Nd2 Kg7 34.Ne4 Kf7 35.Nc5 Ra8 36.Rxb4 Ra7 37.Rd4 Ke8
38.Kc2 a5 39.Kc3 Kd8 40.Ne6+ Kd7 41.Nc5+ Kd8 42.Re4 h6 43.b3 Ra8 44.g3 Rb8 45.Rxe5 fxe5 46.f6 Ke8 47.d7+
Kd8 48.Kd3 Ra8 1-0
16.h7 Qxb1 17.hxg8Q+ Kd7 18.Qxf7 Qxc2+ 19.Kf3 Nc6 20.Qgxe6+ Kc7 21.Qf4+ Kb6 22.Qee3+ Bc5 23.g8Q b1Q
24.Rh6 Qxf1 25.Qb4+ Qb5 26.Qd8+ 1-0
23...Nxh4 24.gxh4 Qxh4 25.Nde7+ Kh8 26.Nxf5 Qh2+ 27.Kf1 Re6 28.Qb7 Rg6 29.Qxa8+ Kh7 30.Qg8+ Kxg8
31.Nce7+ Kh7 32.Nxg6 fxg6 33.Nxg7 Nf2 34.Bxf4 Qxf4 35.Ne6 Qh2 36.Rdb1 Nh3 37.Rb7+ Kh8 38.Rb8+ Qxb8
39.Bxh3 Qg3 0-1
10.b4 c5 11.b5 c4 12.Rb1 d4 13.Ne4 axb5 14.0-0 Rxa2 15.d3 Rxc2 16.Bd1 Ra2 17.f5 Nxe5 18.fxe6 f6 19.Rxf6 gxf6
20.Nxf6+ Kd8 21.Nd5 Qd6 22.Bg5+ Kc8 23.Bg4 Nxg4 24.e7 Bxe7 25.Qxg4+ Nd7 26.Nxe7+ Kc7 27.Bf4 Ne5
28.Qg7 Kb6 29.Bxe5 Qf6 30.Bxd4+ 1-0
15.Bf6 Rb8 16.Rad1 gxf6 17.exf6 b5 18.Ncxb5 axb5 19.Nxb5 Rxb5 20.Bxb5 Qb7 21.c4 Be4 22.Qe2 Rg8 23.f3 dxc4
24.fxe7 Kxe7 25.Qxc4 Bd5 26.Rxd5 exd5 27.Qxc6 Qxc6 28.Bxc6 Kd6 29.Bb7 Bg7 30.Rd1 Bxb2 31.Bxd5 Ke7
32.Rb1 Bd4+ 33.Kf1 Rc8 34.Rb7+ Kd6 35.Bb3 Rc1+ 36.Ke2 f6 37.Rh7 Rg1 38.g3 Rg2+ 39.Kd3 Bg1 40.Rxh6 Ke5
41.f4+ Kf5 42.Ba4 Ke6 43.Ke4 Bxh2 44.Bb3+ Ke7 45.Kf5 Bxg3 46.Rxf6 Rf2 47.Re6+ Kd7 48.Re4 Bh4 49.Kg6
Be7 50.Kf7 Ba3 51.Be6+ Kd6 52.Rd4+ Kc5 53.Ra4 Bb2 54.f5 Rh2 55.f6 Rh6 56.Rf4 Kd6 57.Bb3 Bc3 58.Rf1 Bd4
59.Rf5 Be5 60.Rf3 Bd4 61.a4 Bb2 62.Bc4 Kc5 63.Bd3 Kb4 64.Bb5 Rh1 65.Ke6 Re1+ 66.Kd7 Rd1+ 67.Kc8 Rc1+
68.Kb7 1-0