Game Theory Exercises 2

Game Theory - Exercises
1. Eliminating Dominated Strategies

Test your knowledge on the following questions.
a) Solve the normal form game in of Fig. 1. by eliminating dominated strategies. Verify that the
resulting solution is a Nash equilibrium of the game.
N C J
N 73,25 57,42 66,32
C 80,26 35,12 32,54
J 28,27 63,31 54,29
Figure 1 Eliminating dominated strategies in a 3 x 3 normal form game.
b) Can a Nash equilibrium to the game in Fig. 2 be found by the iterated elimination of dominated
strategies? If so, describe exactly in what order you delete strategies.
a b c d e
A 63, -1 28, -1 -2, 0 -2, 45 -3, 19
B 32, 1 2, 2 2, 5 33, 0 2, 3
C 54, 1 95, -1 0, 2 4, -1 0, 4
D 1, -33 -3, 43 -1, 39 1, -12 -1, 17
E -22, 0 1, -13 -1, 88 -2, -57 -3, 72
Figure 2 Eliminating dominated strategies in a 5 x 5 normal form game.
2. Second- Price Auction
A single object is to be sold at auction. There are n > 1 bidders, each submitting a single bid,
in secret, to the seller. The value of the object to bidder i is v
i
. The winner of the object is the
highest bidder, but i pays only the next highest bid.
a) Show that truth-telling (i.e., each player i bids v
i
) is a dominant strategy for each player. To
simplify the argument, you can assume, where convenient, that there are no ties.
b) Does your analysis depend on whether or not others tell the truth?
c) What are some reasons that real second-price auctions might not conform to the assumptions
of this model?
3. An Increasing-Bid Auction
A Ming vase is to be sold at auction. There are n bidders, and the value of the vase to bidder
i is v
i
> 0, i = 1, ...., n. The auctioneer begins the bidding at zero, and raises the price at the rate
of $ 1 per second. All bidders who are willing to buy the vase at the stated price put their hands
up simultaneously.This last and highest bidder is the winner of the auction and must buy the vase
for the stated price. Show that this auction has the same optimal strategies and the same outcome
as the Second-Price Auction.
HINT: First show that the only undominated pure strategies for bidder i take the form of
choosing a price p
i
and keeping a hand up until the auctioneers price goes higher than this. Then
nd the optimal p
i
.
1
4. Eliminating Dominated Strategies adAbsurdum
Consider an n-player game in which each player announces simultaneously an integer between
1 and 1000. The winner is the player whose announcement is closed to 1/2 of the average of all
the announcements. In the case of a tie, the price is given by a random draw among the winners.
a) Show that if players iterate the elimination of dominated strategies, the only remaining strategy
is to announce 1.
b) To most people, this does not sound like what would actually happen if n people play this game,
probably because people do, at most, a few iterations of dominated strategies. How many levels
of iteration do you think people actually engage in?
c) Extra credit: Stage the game with some friends, with the payo such that the winner receives
$10.00. Can you estimate how many stage of elimination dominated strategies people will go
through?
5. A Pure Coordination Game
Three people independently choose an integer between two and nine. If the three choices are
the same, each person receives the amount chosen. Otherwise each person loses the amount the
person chose.
a) What are the pure strategy Nash equilibria of this game?
b) How do you think people will actually play this game?
c) What doses the game look like if you allow communication among the players before they make
their choices? How would you model such communication, and how do you think communication
would change the behavior of the players?
6.Variations on Duopoly
In a certain market there are two rms, which we label a and b. If the rms produce output
q
a
and q
b
, then the price they will receive for their goods is given by p = (q
a
+ q
b
) or zero
if p would otherwise be negative. Each rm has marginal cost c > 0 and no xed costs. Suppose
> 3c (youll see why we make this assumption as you go through the problem.)
a) Suppose the rm choose the quantity q
a
and q
b
independently in each period, without regard
to their behavior in previous periods, and each maximizes prots
a
= (p c)q
a
and
b
=
(p c)q
b
.Find the unique pure strategy Nash equilibrium to this game. This is called the
CournotDuopoly model.
b) Suppose the two rms collude by agreeing that each will produce an amount q
= q
a
= q
b
,
and they have some way to enforce the agreement. What should they choose for q
? What
are the prots of the two rms? Compare this with the Cournot duopoly prots. This Nash
equilibrium is called the Monopolymodel or the Cartelmodel.
c) Suppose rm a reneges on its promise on the previous part but b doses not. What should a
choose for q
a
? What are as prots, and what are bs prots?
d) Suppose rm b nds out that rm a is going to do in the previous part, and chooses q
b
to
maximize prots, given what rm a is going to do. What is q
b
now? What do you think
happens if they go back and forth this way forever?
2
e) Suppose the two rms choose price as opposed to quantity, where costumers all go to the lowest-
price rm, and the rm split the market if they choose the same price. What is the unique Nash
equilibrium of this game? This is called the Bertrandduodolymodel. Bertands result is often
called a paradox. In what sense do you think this is an accurate description of this solution?
How do you explain the dramatic dierence in outcomes between this and the Cournot model?
f) Suppose rm a gets to choose its output rst and only afterwards doses rm b get to choose its
output. Find the equilibrium choices of q
a
and q
b
in this case, and compare the prots of the
two rms with the Cournot duopoly case. This is called the Stackelbergduopolymodel. HINT:
Firm a should nd its prot for every q
a
, given that rm b will choose q
b
to maximize its prots
given q
a
. Then, among all these prots, rm a choose q
a
to maximize prots.
7. The Rotten Kid Theorem
This problem is core of Gary Beckers (1981) famous theory of the family. You might check the
original, through, since Im not sure I got the genders right.
1
A certain family consists of a mother and a son, with increasing, concave utility funktions u(y)
for the mother and v(z) for the sun. The sun can aect both his income and of the mother by
choosing a level of familial work commitment a, so y = y(a) and z = z(a). The mother, however,
feels a degree of altruism > 0 towards the son, so given y and z, she transfers an amount t to
the son to maximize the objective function.
u(y t) + v(z + t) (1)
The son, however is perfectly selsh (rotten), and chooses the level of a to maximize his own
utility v(z(a) + t). However he knows that his mothers transfer depends on y and z, and hence
on a.
a) Use backward induction to show, what the son chooses a to maximize total family income
y(a) + z(a).
b) Show that t is an increasing function of . Also if we write y = y(a) + y, and t is an increasing
function of the mothers exogenous wealth y.
c) Show that for suciently small > 0, t < 0; i.e., the transfer is from the son to the mother.
HINT: For each a, the mother observes y(a) and z(a), and chooses t = t(a) to maximize (1).
This gives the rst-order condition u
+v
= 0.Dierentiate the rst-order condition with respect

to a to show that the solution to the rst-order condition for the maximization of z(a) + t(a) is
the same as for the maximization of y(a) + z(a) = 0.
1
The Rotten Kid Theorem has been empirically tested and receives some support, through
the evidence falls short (often way short) of conrming that cross-generational families maximize
joint income. See Cox 1987, 1990; Cox and Rank 1992; and Altonji et al. 1992, 1997.
3

Game Theory Exercises 2

Uploaded by

Copyright:

Available Formats

Game Theory Exercises 2

Uploaded by

Document Information

Original Description:

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

Game Theory Exercises 2

Uploaded by

Copyright:

Available Formats

Game Theory - Exercises

1. Eliminating Dominated Strategies

= 0.Dierentiate the rst-order condition with respect

You might also like