Microeconomics: Decision Theory, MQuEA Prof.

Arguedas

Review exercises, Chapter #3

_____________________________________________________________________________

GAMES OF COMPLETE INFORMATION: Exercises for review

1. Consider the following normal-form game:

PLAYER 2
L C R
U 2, 0 1, 1 4, 2
PLAYER 1 M 3, 4 1, 2 2, 3
D 1, 3 0, 2 3, 0

a) What strategies survive after the elimination of strictly dominated strategies?

b) Find all the pure and mixed strategy Nash equilibria of this game.

Solution:

a)

D is strictly dominated by U, thus C is strictly dominated by R. The strategies that survive are

PLAYER 2
L R
U 2, 0 4, 2
PLAYER 1 M 3, 4 2, 3

b)

PLAYER 2 1’s
L (p) R (1-p) Expected
Payoff
U (q) 2, 0 4, 2 2p+(1-p)4
PLAYER 1
M (1-q) 3, 4 2, 3 3p+(1-p)2
2’s Expected Payoff (1-q)4 2q+(1-q)3

2 Pure Strategy Nash Equilibria = (M,L), (U,R) and 1 Mixed Strategy Nash Equilibria

2𝑝 1 𝑝 4 3𝑝 1 𝑝 2⇒ 3𝑝 2⇒𝑝 2
3

Player 1 will be willing to randomize what to play if player 2 chooses L 2/3 of the time, chooses U
if player 2 chooses L less than 2/3 of the time, and chooses M if player 2 chooses R less than 1/3
of the time.
Microeconomics: Decision Theory, MQuEA Prof. Arguedas

Review exercises, Chapter #3

_____________________________________________________________________________

1 𝑞 4 2𝑞 1 𝑞 3⇒ 3𝑞 1⇒𝑞 1/3

Player 2 will be willing to randomize what to pay if player 1 chooses U 1/3 of the time, chooses L
if player 1 chooses U less than 1/3 of the time and chooses R if player 1 chooses M less than 2/3
of the time.

Mixed strategy probabilities:

PLAYER 2
L R
U 2/3*1/3=2/9 1/3*1/3=1/9
PLAYER 1 M 2/3*2/3=4/9 1/3*2/3=2/9

The most likely outcome is 𝐿, 𝑀 .

2. Repeat the same exercise with the following normal-form game:

PLAYER 2
L C R
U 3, 1 0, 0 5, 0
PLAYER 1 M 2, 1 1, 2 3, 1
D 1, 2 0, 1 4, 4

Solution:

a)

D is strictly dominated by U, thus C strictly dominates R. The strategies that survive are

PLAYER 2
L C
U 3, 1 0, 0
PLAYER 1 M 2, 1 1, 2

b)

PLAYER 2
Microeconomics: Decision Theory, MQuEA Prof. Arguedas

Review exercises, Chapter #3

_____________________________________________________________________________

L (p) C (1-p) 1’s

Expected
Payoff
PLAYER 1
U (q) 3, 1 0, 0 3p
M (1-q) 2, 1 1, 2 2p+(1-p)1
2’s Expected Payoff 1 (1-q)2

2 Pure Strategy Nash Equilibria = (U,L), (M,C) and 1 Mixed Strategy Nash Equilibria

3𝑝 2𝑝 1 𝑝 1 ⇒ 2𝑝 1⇒𝑝 1
2

Player 1 will be willing to randomize what to play if player 2 chooses L 1/2 of the time, chooses U
if player 2 chooses L more than ½ of the time and chooses M if player 2 chooses C more than ½
of the time.

1 1 𝑞 2 ⇒ 2𝑞 1⇒𝑞 1
2

Player 2 will be willing to randomize what to pay if player 1 chooses U 1/2 of the time, chooses L
if player 1 chooses U more than ½ of the time and chooses C if player 1 chooses M more than ½
of the time.

Mixed strategy probabilities:

PLAYER 2
L C
U 1/2*1/2=1/4 1/2*1/2=1/4
PLAYER 1 M 1/2*1/2=1/4 1/2*1/2=1/4

All outcomes are equally likely.

3. The Game of Chicken is played by two macho teens who speed towards each other on a single-
lane road. The first to veer is branded the chicken, whereas the one who does not veer gains per-
group esteem. Of course, if neither veers, both die in the resulting crash. Payoffs to the Chicken
game are provided in the following table:

TEEN 2
Veer Not veer
TEEN 1 Veer 2, 2 1, 3
Not veer 3, 1 0, 0
Microeconomics: Decision Theory, MQuEA Prof. Arguedas

Review exercises, Chapter #3

_____________________________________________________________________________

a) Find all the pure strategy Nash equilibria.

b) Find the mixed strategy Nash equilibrium. What is the probability that both teenagers will
survive?
c) Suppose the game is played sequentially with teen 1 moving first and committing to this
action by throwing away the steering wheel. Write down the extensive form game and
find the corresponding sub-game perfect equilibrium.
d) Does the equilibrium found in part c) coincide with any of the equilibria found in parts a)
or b)? Why?

a)

2 Pure Strategy Nash Equilibria = (NV,V), (V,NV)

b)

TEEN 2 1’s
V (p) NV (1-p) Expected
Payoff
V (q) 2, 2 1, 3 2p+(1-p)1
TEEN 1
NV(1-q) 3, 1 0, 0 3p
2’s Expected Payoff 2q+(1-q)1 3q

2𝑝 1 𝑝 1 3𝑝 ⇒ 2𝑝 1⇒𝑝 1
2

Teen 1 will be willing to randomize what to play if teen 2 chooses NV 1/2 of the time, chooses V
if teen 2 plays V less than ½ of the time and chooses NV if teen 1 plays NV less than ½ of the
time.

2𝑞 1 𝑞 1 3𝑞 ⇒ 2𝑞 1⇒𝑞 1
2

Teen 2 will be willing to randomize what to play if teen 1 chooses V 1/2 of the time, chooses V if
teen 1 plays V less than ½ of the time and chooses NV if teen 1 plays NV less than ½ of the time.

Mixed strategy probabilities:

PLAYER 2
V NV
V 1/2*1/2=1/4 1/2*1/2=1/4
PLAYER 1 NV 1/2*1/2=1/4 1/2*1/2=1/4
Microeconomics: Decision Theory, MQuEA Prof. Arguedas

Review exercises, Chapter #3

_____________________________________________________________________________

The probability that both teenagers survive is P(V,V) + P(V,NV) + P(NV,V)= 1/4+1/4+1/4=3/4.

c)

We can use a decision tree and backwards induction:

(2,2)
V
2
(1,3)
NV
V
1
NV

2 V (3,1)

NV (0,0)

The sub-game perfect equilibrium is (NV,V)

d)

It is one of the pure strategy nash equilibria found in part a), and it was one of the equally probable
results of the mixed strategy.

4. Consider the following normal‐form game:

PLAYER 2
A B
PLAYER 1 A 0, 0 3, ‐1
B ‐1, 3 1, 1

a) Verify that the Nash equilibrium is the usual one for the Prisoners’ Dilemma and that
both players have dominant strategies.
b) Suppose the stage game is repeated infinitely many times. Compute the discount factor
required for the players to sustain the cooperative strategy (B,B). Outline the trigger
strategies you are considering.

a)
Microeconomics: Decision Theory, MQuEA Prof. Arguedas

Review exercises, Chapter #3

_____________________________________________________________________________

B of player 2 is strictly dominated by A of player 2 and B of player 1 is dominated by A of player

1, delivering a Nash equilibrium= (A,A), which is inefficient (both players are better off by
selecting (B,B), but this is not an equilibrium).

b)

The objective is (B,B)=(1,1). The trigger strategy might be a grim strategy in which cooperation
exists in round 𝑡 and in subsequent rounds as long as the other player does not defect from the
agreement. With these outcomes, cooperation would translate in a Net Present Value 𝑁𝑃𝑉

1 𝛿 𝛿 ⋯ with a discount factor 𝛿 and defect would have 𝑁𝑃𝑉 3 𝛿∙0

𝛿 ∙0 ⋯ 3. Cooperation will be attained and sustained if 𝑁𝑃𝑉 𝑁𝑃𝑉 , which is

achieved with 3⇒ 2 3𝛿 ⇒ 𝛿 .

6. Consider the classical Prisoners’ Dilemma game, where the non‐cooperative outcome is the
unique Nash equilibrium of the stage game.

a) Explain this game. Invent numbers for the different outcomes, represent them in a bi‐
matrix and show how the Nash‐equilibrium is achieved.
b) Suppose that your invented game is repeated infinitely many periods. Explore the
possibility that players can support the cooperative outcome by using tit‐for‐tat
strategies. How many periods of punishment would be enough to sustain cooperation?
c) How would the results change if players instead used grim strategies to sustain
cooperation?

a)

The dilemma is the following: To cooperate with their partner for mutual reward, or betray their
partner for individual reward. In the invented game, rewards are in the form of higher payoffs.

Prisoner 2
ND D
Prisoner 1 ND 3, 3 1, 4
D 4, 1 2, 2
Microeconomics: Decision Theory, MQuEA Prof. Arguedas

Review exercises, Chapter #3

_____________________________________________________________________________

Where D=To declare, ND=Not to declare. Nash Equilibrium=(D,D).

b)
In a tit‐for‐tat strategy, players mimic the action of the opponent. The objective is (ND,
ND)=(3,3). Cooperation would translate in a Net Present Value 𝑁𝑃𝑉 3 3𝛿 3𝛿 ⋯

with a discount factor 𝛿. The number of periods of punishment to sustain cooperation will

directly depend on 𝛿, for example, alternate between outcomes where Prisoner 1 cooperates
and Prisoner 2 deviates, and vice versa, then defect would have 𝑁𝑃𝑉 4 𝛿 4𝛿 𝛿

⋯ . Cooperation will be attained and sustained if 𝑁𝑃𝑉 𝑁𝑃𝑉 , which is

∙ ∙ ∙
achieved with ⇒ 4 𝛿⇒ 4 𝛿⇒3 3𝛿 4 𝛿⇒

2𝛿 1⇒𝛿 . Therefore, with these payoffs and if 𝛿 , just one period of punishment

would be enough to sustain cooperation.

c)

Under grim strategy, cooperation exists in the first round and in subsequent rounds as long as
the other player does not defect from the agreement. With these outcomes, 𝑁𝑃𝑉 is like in b)

and defect would have 𝑁𝑃𝑉 4 𝛿 𝛿 ⋯ 4 . Cooperation will be attained and

sustained if 𝑁𝑃𝑉 𝑁𝑃𝑉 , which is achieved with 4 ⇒3 4 4𝛿 𝛿 ⇒ 3𝛿

1⇒𝛿 . Therefore, grim strategy works for a smaller 𝛿 than the tit‐for‐tat strategy proposed

in b), which is reasonable given that grim strategy is more aggressive.