Extensive Form Games: Perfect Information: ECON2112

Extensive Form Games: Perfect Information
ECON2112

Introduction
In a normal form game, players take decisions simultaneously.
In an extensive form game, players may observe other players

past decisions.

Introduction
A game of perfect information is an extensive form game where

players can always observe other players past actions

Perfect Information
Example (Perfect Information)

1
L
2
a
2, 1
1, 3
0, 0
The open circle is the initial node or root of the game tree.
The black circles and the open circle are called decision nodes.
L, R, a and b are called choices.
1 and 2 are the names of the players.
We specify a payoff vector in each final node.

Perfect Information

1
L
2
a
2, 1
1, 3
0, 0

Perfect Information

1
L
2
a
2, 1
1, 3
0, 0

Perfect Information

1
L
2
a
2, 1
1, 3
0, 0

Perfect Information

1
L
2
a
2, 1
1, 3
0, 0

Perfect Information

1
L
2
a
2, 1
2
b
0, 0
c
4, 0
d
1, 3
L, R, a, b, c, d are called choices (or actions).
What is player 1s set of pure strategies? S1 = {L, R }
What is player 2s set of pure strategies? S2 = {ac , ad , bc , bd }

Perfect Information

1
L
2
a
2, 1
2
b
0, 0
c
4, 0
d
1, 3

Perfect Information

1
L
2
a
2, 1
2
b
0, 0
c
4, 0
d
1, 3

Perfect Information

1
L
2
a
2, 1
2
b
0, 0
c
4, 0
d
1, 3

Perfect Information

1
L
2
a
2
b
1
2, 1
0, 0
4, 0
1, 4
1, 3
Player 1 set of pure strategies: S1 = {LA, LB , RA, RB }
Player 2 set of pure strategies: S2 = {ac , ad , bc , bd }.

Perfect Information

1
L
2
a
2
b
1
2, 1
0, 0
4, 0
1, 4
1, 3
Player 1 set of pure strategies: S1 = {LA, LB , RA, RB }

Perfect Information

1
L
2
a
1
2
1
2
2
2, 1
0, 0
4, 0
2, 3
0, 3
We can introduce moves of Nature.
Player 1 set of pure strategies: S1 = {L, R }.

Perfect Information

1
L
2
a
1
2
1
2
2
2, 1
0, 0
4, 0
2, 3
0, 3
We can introduce moves of Nature.
Player 1 set of pure strategies: S1 = {L, R }.

Pure and Mixed Strategies
What is a pure strategy in an Extensive Form Game?

In a game of perfect information, a pure strategy is a plan of action for
the entire game, that tells the player what particular choice to take at
every possible juncture of the game where it is his turn to move.
What is the set of Pure Strategies?

The set of all possible plans of action.
What is a Mixed Strategy?

A probability distribution on the set of pure strategies.









Normal Form Representation
Once we know what is the set of pure strategies for each player,
we can obtain the normal form representation of a an extensive
form game
Example
1
L
2
a
2, 1
b
0, 0
1, 3
L
R
a
2, 1
1, 3
b
0, 0
1, 3

form game
Example
1
L
2
a
2, 1
b
0, 0
1, 3
L
R
a
2, 1
1, 3
b
0, 0
1, 3

form game
Example
1
L
2
a
2, 1
b
0, 0
1, 3
L
R
a
2, 1
1, 3
b
0, 0
1, 3

form game
Example
1
L
2
a
2, 1
b
0, 0
1, 3
L
R
a
2, 1
1, 3
b
0, 0
1, 3

form game
Example
1
L
2
a
2, 1
b
0, 0
1, 3
L
R
a
2, 1
1, 3
b
0, 0
1, 3

Example
1
L
2
a
2, 1
2
b
0, 0
c
4, 0
d
1, 3
L
R
ac
2, 1
4, 0
ad
2, 1
1, 3
bc
0, 0
4, 0
bd
0, 0
1, 3

Example
1
L
2
a
2, 1
2
b
0, 0
c
4, 0
d
1, 3
L
R
ac
2, 1
4, 0
ad
2, 1
1, 3
bc
0, 0
4, 0
bd
0, 0
1, 3

Example
1
L
2
a
2, 1
2
b
0, 0
c
4, 0
d
1, 3
L
R
ac
2, 1
4, 0
ad
2, 1
1, 3
bc
0, 0
4, 0
bd
0, 0
1, 3

Example
1
L
2
a
2, 1
2
b
0, 0
c
4, 0
d
1, 3
L
R
ac
2, 1
4, 0
ad
2, 1
1, 3
bc
0, 0
4, 0
bd
0, 0
1, 3

Example
1
L
2
a
2, 1
2
b
0, 0
c
4, 0
d
1, 3
L
R
ac
2, 1
4, 0
ad
2, 1
1, 3
bc
0, 0
4, 0
bd
0, 0
1, 3

Example
1
L
2
a
2, 1
2
b
0, 0
c
4, 0
d
1, 3
L
R
ac
2, 1
4, 0
ad
2, 1
1, 3
bc
0, 0
4, 0
bd
0, 0
1, 3

Normal From Representation
Example
1
L
2
a
2
b
1
2, 1
0, 0
4, 0
1, 4
1, 3
LA
LB
RA
RB
ac
2, 1
2, 1
4, 0
1, 4
ad
2, 1
2, 1
1, 3
1, 3
bc
0, 0
0, 0
4, 0
1, 4
bd
0, 0
0, 0
1, 3
1, 3

Example
1
L
2
a
2
b
1
2, 1
0, 0
4, 0
1, 4
1, 3
LA
LB
RA
RB
ac
2, 1
2, 1
4, 0
1, 4
ad
2, 1
2, 1
1, 3
1, 3
bc
0, 0
0, 0
4, 0
1, 4
bd
0, 0
0, 0
1, 3
1, 3

Example
1
L
2
a
2
b
1
2, 1
0, 0
4, 0
1, 4
1, 3
LA
LB
RA
RB
ac
2, 1
2, 1
4, 0
1, 4
ad
2, 1
2, 1
1, 3
1, 3
bc
0, 0
0, 0
4, 0
1, 4
bd
0, 0
0, 0
1, 3
1, 3

Example
1
L
2
a
2
b
1
2, 1
0, 0
4, 0
1, 4
1, 3
LA
LB
RA
RB
ac
2, 1
2, 1
4, 0
1, 4
ad
2, 1
2, 1
1, 3
1, 3
bc
0, 0
0, 0
4, 0
1, 4
bd
0, 0
0, 0
1, 3
1, 3

Example
1
L
2
a
2
b
1
2, 1
0, 0
4, 0
1, 4
1, 3
LA
LB
RA
RB
ac
2, 1
2, 1
4, 0
1, 4
ad
2, 1
2, 1
1, 3
1, 3
bc
0, 0
0, 0
4, 0
1, 4
bd
0, 0
0, 0
1, 3
1, 3

Example
1
L
2
a
2
b
1
2, 1
0, 0
4, 0
1, 4
1, 3
LA
LB
RA
RB
ac
2, 1
2, 1
4, 0
1, 4
ad
2, 1
2, 1
1, 3
1, 3
bc
0, 0
0, 0
4, 0
1, 4
bd
0, 0
0, 0
1, 3
1, 3

Example
1
L
2
a
2
b
1
2, 1
0, 0
4, 0
1, 4
1, 3
LA
LB
RA
RB
ac
2, 1
2, 1
4, 0
1, 4
ad
2, 1
2, 1
1, 3
1, 3
bc
0, 0
0, 0
4, 0
1, 4
bd
0, 0
0, 0
1, 3
1, 3

Perfect Information
Example
1
L
2
a
1
2
1
2
2
2, 1
0, 0
4, 0
2, 3
0, 3
L
R
ac
2, 1
2, 23
ad
2, 1
1, 3
bc
0, 0
2, 32
bd
0, 0
1, 3

Perfect Information
Example
1
L
2
a
1
2
1
2
2
2, 1
0, 0
4, 0
2, 3
0, 3
L
R
ac
2, 1
2, 23
ad
2, 1
1, 3
bc
0, 0
2, 32
bd
0, 0
1, 3

Perfect Information
Example
1
L
2
a
1
2
1
2
2
2, 1
0, 0
4, 0
2, 3
0, 3
L
R
ac
2, 1
2, 23
ad
2, 1
1, 3
bc
0, 0
2, 32
bd
0, 0
1, 3

Perfect Information
Example
1
L
2
a
1
2
1
2
2
2, 1
0, 0
4, 0
2, 3
0, 3
L
R
ac
2, 1
2, 23
ad
2, 1
1, 3
bc
0, 0
2, 32
bd
0, 0
1, 3

Perfect Information
Example
1
L
2
a
1
2
1
2
2
2, 1
0, 0
4, 0
2, 3
0, 3
L
R
ac
2, 1
2, 23
ad
2, 1
1, 3
bc
0, 0
2, 32
bd
0, 0
1, 3

Perfect Information
Example
1
L
2
a
1
2
1
2
2
2, 1
0, 0
4, 0
2, 3
0, 3
L
R
ac
2, 1
2, 23
ad
2, 1
1, 3
bc
0, 0
2, 32
bd
0, 0
1, 3

The normal form representation of an extensive form game is

useful to
Compute Nash equilibria
Find Dominated Strategies.

Nas Equilibrium
Definition (Nash Equilibrium)

The Nash equilibrium of a perfect information game (or, more
generally, of an extensive form game) is the Nash equilibrium of its
Normal form representation.

Nash Equilibrium
Typically, not every Nash equilibrium of a perfect information game is

reasonable.
Example
1
L
2
a
2, 1
1, 3
L
R
a
2, 1
1, 3
b
0, 0
1, 3
0, 0
The game has two Nash equilibria (L, a) and (R , b).
Under (R , b), player 2 is threating player 1 with playing b if he

plays L.
But, is this a credible threat?

Nash Equilibrium

reasonable.
Example
1
L
2
a
2, 1
1, 3
L
R
a
2, 1
1, 3
b
0, 0
1, 3
0, 0

plays L.

Nash Equilibrium

reasonable.
Example
1
L
2
a
2, 1
1, 3
L
R
a
2, 1
1, 3
b
0, 0
1, 3
0, 0

plays L.

Nash Equilibrium
Example
1
L
2
a
2, 1
1, 3
L
R
a
2, 1
1, 3
b
0, 0
1, 3
0, 0
If player 2 finds himself in a situation where he has to choose

between a and b, he will always choose a.
Player 1 should foresee this, and play L instead of R.
(In any case, note that b is a dominated strategy.)

Backwards Induction
This idea can be generalized as the Principle of Backwards

Induction.
Definition (Backwards Induction/Informal Idea)

Players should make their choices in a way consistent with deductions
about other players rational behavior in the future.
The implementation of the Backwards Induction principle to

games of perfect information involves a computational method.

Backwards Induction. Computation
Backwards Induction. Computation
First we resolve the last moves in the game.
Then we substitute decision nodes by the expected payoff that

arises from a rational move at such decision node.
And we keep resolving last moves in the same way until we get to
the root of the game tree.

Rollback equilibrium
Example
1
L
2
a
2
b
1
2, 1
0, 0
1, 3
4, 0 1, 4
The optimal moves of the Rollback equilibrium are A, a, d and L.
This corresponds to the pure strategy profile (LA, ad ).
However, we can also give the result as the behavioral strategy

profile (A, a, d , L).

Example
1
L
2
a
2, 1
2
b
0, 0
c
4, 0
d
1, 3


Example
1
L
R
2
2, 1
c
4, 0
d
1, 3


Example
1
L
2, 1
R
1, 3


Example
1
L
2, 1
R
1, 3


Example
1
L
2, 1
R
1, 3


Example
1
L
2, 1
R
1, 3


Example
1
L
2
a
1
2
1
2
2
2, 1
0, 0
0, 3
4, 0 2, 3
Moves that conform with Backwards Induction: d , a, L.
Expressed as a pure strategy profile: (L, ad ).
Expressed as a behavioral strategy profile: (L, a, d ) .

Example
1
L
2
a
2, 1
b
0, 0
R
1
2
2, 3
1
2
0, 3

Example
1
L
2, 1
R
1
2
2, 3
1
2
0, 3

Example
1
L
2, 1
R
1, 3

Example
1
L
2, 1
R
1, 3

Example
1
L
2, 1
R
1, 3

Example
1
L
2, 1
R
1, 3

Behavioral Strategies
Definition (Behavioral Strategies)

In a game of perfect information, a players behavioral strategy
specifies at every decision node where it is his turn to move a
probability distribution over the choices that follow.
Definition (Behavioral Strategy Profile)

In a game of perfect information, a behavioral strategy profile specifies
one behavioral strategy for each player.

Compare:
Definition (Pure strategy)

Definition (Mixed Strategy)

A mixed strategy is a probability distribution over the set of pure
strategies.
Definition (Behavioral Strategies)

In a game of perfect information, a players behavioral strategy
specifies at every decision node where it is his turn to move a
probability distribution over the choices that follow.

Example
1
L
2
a
2
b
1
2, 1
0, 0
4, 0
1, 4
1, 3
LA
LB
RA
RB
ac
2, 1
2, 1
4, 0
1, 4
ad
2, 1
2, 1
1, 3
1, 3
bc
0, 0
0, 0
4, 0
1, 4
(LA, ac ) corresponds to (L, a, A, c ).
( 12 LA + 21 LB , 13 ac + 32 ad ) c.t. (L, 21 A + 12 B , a, 31 c + 23 d ).
bd
0, 0
0, 0
1, 3
1, 3

Theorem (Kuhn)
(In any extensive form game with perfect recall,) every probability
distributions on ending nodes induced by a mixed strategy profile can
also be induced by some behavioral strategy profile.
The consequence is that, although mixed strategies may look

more general, there is no loss of generality in working only with
behavioral strategies.

Extensive Form Games: Perfect Information: ECON2112

Uploaded by

Copyright:

Available Formats

Extensive Form Games: Perfect Information: ECON2112

Uploaded by

Document Information

Original Description:

Original Title

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

Extensive Form Games: Perfect Information: ECON2112

Uploaded by

Copyright:

Available Formats

Extensive Form Games: Perfect Information