University Carlos III of Madrid Departament of Economics

Game Theory: Problem set 1. Solutions.

Problem 1: For each one of the following normal form games find

(a) The pure-strategy Nash equilibria and their payoffs;

(b) The mixed-strategy Nash equilibria and their payoffs;
(c) Are there any equilibria in dominant strategies? If yes, which ones?

C D
(a) A −50, −50 100, 0
B 0, 100 0, 0
Solution: The best responses of the players are

C D
A −50, −50 100, 0
B 0, 100 0, 0

There are two NE in pure strategies: (A, D) and (B, C). We look for a NE in mixed strategies of the
form

σ1 = xA + (1 − x)B
σ2 = yC + (1 − y)D

The expected utilities of the players are

u1 (A, σ2 ) = −50y + 100(1 − y) = 100 − 150y

u1 (B, σ2 ) = 0 × y + 0 × (1 − y) = 0

and

u2 (σ1 , C) = 50x + 100(1 − x) = 100 − 150x

u2 (σ1 , D) = 0 × x + 0 × (1 − x) = 0

Equating the utilities of the players

100 − 150y = 0, 100 − 150x = 0

we obtain x = y = 32 . Thus, there is an additional NE in mixed strategies,

2 1
σ1 = A+ B
3 3
2 1
σ2 = C+ D
3 3

C D
(b) A −1, 2 1, −2
B 1, −2 −1, 2
Solution: The best responses of the players are

C D
A −1, 2 1, −2
B 1, −2 −1, 2

1
 There are no NE in pure strategies. We look for a NE in mixed strategies of the form
σ1 = xA + (1 − x)B
σ2 = yC + (1 − y)D
The expected utilities of the players are
u1 (A, σ2 ) = −y + (1 − y) = 1 − 2y
u1 (B, σ2 ) = y − (1 − y) = 2y − 1

and
u2 (σ1 , C) = 2x − 2(1 − x) = 4x − 2
u2 (σ1 , D) = −2x + 2(1 − x) = 2 − 4x
Equating the utilities of the players
1 − 2y = 2y − 1, 4x − 2 = 2 − 4x
we obtain x = y = 21 . Thus, there is an additional NE in mixed strategies,
1 1
σ1 = A+ B
2 2
1 1
σ2 = C+ D
2 2

C D
(c) A −10, −10 0, −20
B −20, 0 10, 10
Solution: The best responses of the players are

C D
A −10, −10 0, −20
B −20, 0 10, 10

There are two NE in pure strategies: (A, C) and (B, D). We look for a NE in mixed strategies of the
form
σ1 = xA + (1 − x)B
σ2 = yC + (1 − y)D
The expected utilities of the players are
u1 (A, σ2 ) = −10y + 0 × (1 − y) = −10y
u1 (B, σ2 ) = −20y + 10(1 − y) = 10 − 30y

and
u2 (σ1 , C) = −10x + 0 × (1 − x) = −10x
u2 (σ1 , D) = −20x + 10(1 − x) = 10 − 30x
Equating the utilities of the players
−10y = 10 − 30y, −10x = 10 − 30x
we obtain x = y = 21 . Thus, there is an additional NE in mixed strategies,
1 1
σ1 = A+ B
2 2
1 1
σ2 = C+ D
2 2

2
 C D
(d) A 3, 3 0, 0
B 0, 0 1, 1
Solution: We look for a NE of the form

σ1 = xA + (1 − x)B
σ2 = yC + (1 − y)D

The expected utilities of the players are

u1 (σ1 , σ2 ) = 1 − x − y + 4xy = 1 − y + x(4y − 1)

u2 (σ1 , σ2 ) = 1 − x − y + 4xy = 1 − x + y(4x − 1)

We see that 
1
1
 if y > 4
1
BR1 (y) = [0, 1] if y = 4
1

0 if y <

4

and 
1
1
 if x > 4
1
BR2 (x) = [0, 1] if x = 4
1

0 if x <

4

Graphically,

BR2 (x)
1

BR1 (y)

1/4

1/4 1 x

And we obtain three NE

(a) (A, C) with payoffs u1 = u2 = 3.
(b) (B, D) with payoffs u1 = u2 = 1.
(c) 14 A + 34 B, 41 C + 34 D with payoffs u1 = u2 = 43 .

C D
(e) A 4, −8 −2, −7
B 1, −5 −1, −6
Solution: We look for a NE of the form

σ1 = xA + (1 − x)B
σ2 = yC + (1 − y)D

3
 The expected utilities of the players are

u1 (σ1 , σ2 ) = −1 − x + 2y + 4xy = −1 + 2y + x(4y − 1)

u2 (σ1 , σ2 ) = −6 − x + y − 2xy = −6 − x + y(1 − 2x)

We see that 
1
1
 if y > 4
1
BR1 (y) = [0, 1] if y = 4
1

0 if y <

4

and 
1
1
 if x < 2
1
BR2 (x) = [0, 1] if x = 2
1

0 if x >

2

Graphically,

BR2 (x)
1

BR1 (y)

1/4

1/2 1 x

And we obtain a unique NE

 
1 1 1 3
(A, C), (B, D), A + B, C + D
2 2 4 4

with payoffs
1 13
u1 = − , u2 = −
2 2

C D
(f) A 2, 5 8, −1
B 4, −8 −2, −3
Solution: We look for a NE of the form

σ1 = xA + (1 − x)B
σ2 = yC + (1 − y)D

The expected utilities of the players are

u1 (σ1 , σ2 ) = −2 + 10x + 6y − 12xy = −2 + 6y + x(10 − 12y)

u2 (σ1 , σ2 ) = −3 + 2x − 5y + 11xy = −3 + 2x + y(11x − 5)

4
 We see that 
5
1
 if y < 6
5
BR1 (y) = [0, 1] if y = 6
5

0 if y >

6

and 
5
1
 if x > 11
5
BR2 (x) = [0, 1] if x = 11
5

0 if x <

1

Graphically,

BR2 (x)
1
5
6

BR1 (y)

5 x
11 1

And we obtain a unique NE

 
5 6 5 1
(A, C), (B, D), A + B, C + D
11 11 6 6

with payoffs
23
u1 = 3 u2 = −
11

C D
(g) A 10, 10 −2, −7
B 1, −5 1, 1
Solution: We look for a NE of the form

σ1 = xA + (1 − x)B
σ2 = yC + (1 − y)D

The expected utilities of the players are

u1 (σ1 , σ2 ) = 1 − 3x + 12xy = 1 + x(12y − 3)

u2 (σ1 , σ2 ) = 1 − 8x − 6y + 23xy = 1 − 8x + y(23x − 6)

We see that 
1
0
 if y < 4
1
BR1 (y) = [0, 1] if y = 4
1

1 if y >

4

5
 and 
6
1
 if x > 23
6
BR2 (x) = [0, 1] if x = 23
6

0 if x <

23

Graphically,

BR2 (x)
1

BR1 (y)
1
4

6 x
23 1

And we obtain three NE

(a) (A, C) with payoffs u1 = u2 = 10.
(b) (B, D) with payoffs u1 = u2 = 1.
6
A + 17 1 3 25


(c) 23 23 B, 4 C + 4 D with payoffs u1 = 1, u2 = − 23 .

Problem 2: For the normal form games below determine

(a) the strategies that survive the iterated elimination of strictly dominated strategies.
(b) the pure strategy Nash equilibria of the game.
(c) the mixed strategy Nash equilibria of the game.

X Y Z
(a) A 20, 10 10, 20 1, 1
B 10, 20 20, 10 1, 1
C 1, 1 1, 1 0, 0
Solution: Note that C is dominated by A for player 1. Hence, we obtain the game

X Y Z
A 20, 10 10, 20 1, 1
B 10, 20 20, 10 1, 1

Now that Z is dominated by X for player 2. Hence, we obtain the game

X Y
A 20, 10 10, 20
B 10, 20 20, 10

The rationalizable strategies are {A, B} × {X, Y }. The best replies of the players are

6
 X Y
A 20, 10 10, 20
B 10, 20 20, 10

We see that there are no NE in pure strategies. Let us look for a NE in mixed strategies of the form

σ1 = xA + (1 − x)B
σ2 = yX + (1 − y)Y

The expected utilities of player 1 when he chooses pure actions are

u1 (A, σ2 ) = 20y + 10(1 − y) = 10 + 10y

u1 (B, σ2 ) = 10y + 20(1 − y) = 20 − 10y

We must have that 10 + 10y = 20 − 10y. That is y = 21 .

The expected utilities of player 2 when he chooses pure actions are

u2 (σ1 , X) = 10x + 20(1 − x) = 20 − 10x

u2 (σ1 , Y ) = 20x + 10(1 − x) = 10 − 20x

Equating 20 − 10x = 10 − 20x, we obtain x = 12 . We conclude that there is a NE in mixed strategies,

1 1
σ1 = A+ B
2 2
1 1
σ2 = X+ Y
2 2

D E F
(b) A 2, 0 1, 1 4, 2
B 3, 4 1, 2 2, 3
C 1, 3 0, 2 3, 0
Solution: Note that C is dominated by A for player 1,

D E F
A 2, 0 1, 1 4, 2
B 3, 4 1, 2 2, 3

And now E is dominated by F for player 2,

D F
A 2, 0 4, 2
B 3, 4 2, 3

The rationalizable strategies are {A, B} × {D, F }. The best responses are

D F
A 2, 0 4, 2
B 3, 4 2, 3

and we obtain that the NE in pure strategies are (B, D) and (A, F ).

We look now for a NE in mixed strategies of the form

σ1 = xA + (1 − x)B
σ2 = yD + (1 − x)F

The expected utilities of player 1 when he chooses pure actions are

u1 (A, σ2 ) = 2y + 4(1 − y) = 4 − 2y
u1 (B, σ2 ) = 3y + 2(1 − y) = 2 + y

7
 We must have that player 1 is indifferent between A, B. That is 4 − 2y = 2 + y. Therefore, y = 2/3.
The expected utilities of player 2 when he chooses pure actions are

u2 (σ1 , D) = 0x + 4(1 − x) = 4 − 4x
u2 (σ1 , F ) = 2x + 3(1 − x) = 3 − x

We must have that player 2 is indifferent between D and F . That is, 4 − 4x = 3 − x. Therefore,
x = 1/3. We obtain the NE  
1 2 2 1
A + B, D + F
3 3 3 3
with expected payoffs u1 = u2 = 38 .

D E F
(c) A 2, 2 1, 1 4, 3
B 1, 4 0, 5 3, 3
C 3, 3 −2, 2 3, 0
Solution: Note that B is dominated by A for player 1,

D E F
A 2, 2 1, 1 4, 3
C 3, 3 −2, 2 3, 0

And now E is dominated by D for player 2,

D F
A 2, 2 4, 3
C 3, 3 3, 0

The rationalizable strategies are {A, C} × {D, F }. The best responses are

D F
A 2, 2 4, 3
C 3, 3 3, 0

and we obtain that the NE in pure strategies are (C, D) and (A, F ).

We look now for a NE in mixed strategies of the form

σ1 = xA + (1 − x)C
σ2 = yD + (1 − x)F

The expected utilities of player 1 when he chooses pure actions are

u1 (A, σ2 ) = 2y + 4(1 − y) = 4 − 2y
u1 (C, σ2 ) = 3y + 3(1 − y) = 3

We must have that player 1 is indifferent between A and C. That is 4 − 2y = 3. Therefore, y = 21 .

The expected utilities of player 2 when he chooses pure actions are

u2 (σ1 , D) = 2x + 3(1 − x) = 3 − x
u2 (σ1 , F ) = 3x + 0(1 − x) = 3x

We must have that player 2 is indifferent between D and F . That is, 3 − x = 3x. Therefore, x = 43 .
We obtain the NE  
3 1 1 1
A + B, D + F
4 4 2 2
with expected payoffs u1 = 3, u2 = 94 .

8
 D E F
(d) A 4, −1 3, 4 2, 0
B 1, 4 −3, 5 −3, −3
C 3, 3 2, −2 8, −3
Solution: Note that B is dominated by A for player 1,

D E F
A 4, −1 3, 4 2, 0
C 3, 3 2, −2 8, −3

and F is dominated by E for player 2,

D E
A 4, −1 3, 4
C 3, 3 2, −2

Now, C is dominated by A for player 1,

D E
A 4, −1 3, 4

And D is dominated by E for player 2. The set of rationalizable strategies is {A} × {E}. The unique
NE is (A, E).

D E F
(e) A 2, −1 2, −1 4, −2
B 2, −1 −2, −1 3, −2
C 1, 3 −2, 2 3, 4
Solution: Note that C is dominated by A for player 1,

D E F
A 2, −1 2, −1 4, −2
B 2, −1 −2, −1 3, −2

and F is dominated by D for player 2,

D E
A 2, −1 2, −1
B 2, −1 −2, −1

The best responses of the players are

D E
A 2, −1 2, −1
B 2, −1 −2, −1

We have three NE in pure strategies: (A, D), (A, E) and (B, D). We look now for a NE in mixed
strategies of the form

σ1 = xA + (1 − x)B
σ2 = yD + (1 − y)E

The expected utilities of player 1 when he chooses pure actions are

u1 (A, σ2 ) = 2y + 2(1 − y) = 2
u1 (B, σ2 ) = 2y − 2(1 − y) = −2 + 4y

9
 Now, for 0 ≤ y < 1 we have that u1 (A, σ2 ) > −2 + 4y. Hence, in a mixed NE, player 1 will not
choose strategy B with positive probability unless y = 1. The expected utilities of player 2 when he
chooses pure actions are

u2 (σ1 , D) = −x − (1 − x) = −1
u2 (σ1 , E) = −x + −(1 − x) = −1

for any 0 ≤ x ≤ 1. Player 2 is indifferent among all his strategies. We obtain infinitely many NE of
the form,

σ1 = xA + (1 − x)B, 0≤x≤1
σ2 = D

with payoffs u1 = 2, u2 = −1. And we also obtain infinitely many NE of the form,

σ1 = A
σ2 = yD + (1 − y)E, 0≤y≤1

with payoffs u1 = 2, u2 = −1.

A B C D
(f) X 1, 4 2, 2 4, 1 2, 0
Y 0, 0 1, 3 2, 5 0, 0
Z 0, 1 5, 0 8, 1 3, 3
Solution: Note that Y is dominated by X for player 1,

A B C D
X 1, 4 2, 2 4, 1 2, 0
Z 0, 1 5, 0 8, 1 3, 3

Now, B is dominated by A for player 2,

A C D
X 1, 4 4, 1 2, 0
Z 0, 1 8, 1 3, 3

For the moment, we consider the strategies S1 {X, Y }, S2 = {A, C, D}. The best responses of the
player are

A C D
X 1, 4 4, 1 2, 0
Z 0, 1 8, 1 3, 3

We obtain two pure strategy NE: (X, A) and (Z, D). We look now for a NE in mixed strategies of
the form

σ1 = pX + (1 − p)Z
σ2 = xA + yC + (1 − x − y)D

The expected utilities of player 2 when he chooses pure actions are

u2 (σ1 , A) = 4p + (1 − p) = 1 + 3p
u2 (σ1 , C) = p + (1 − p) = 1
u2 (σ1 , D) = 0p + 3(1 − p) = 3 − 3p

We represent this three functions graphically,

10
 3 u2 (σ1 , A) = 1

u2 (σ1 , D) = 3 − 3p

1
u2 (σ1 , C) = 1

1 p
3 1

We see in this picture, that player 2 does not want to consider strategy C. In fact, consider now the
mixed strategy
1 1
σ2 = A + D
2 2
for player 2. Note that

u2 (X, σ2 ) = 2 > 1 = u2 (X, C)

u2 (Z, σ2 ) = 2 > 1 = u2 (Z, C)

That is, strategy C is (strictly) dominated the the mixed strategy σ2 . We may eliminate strategy C.
The rationalizable strategies are {X, Y } × {A, D}. The best responses of the player are

A D
X 1, 4 2, 0
Z 0, 1 3, 3

We obtain two pure strategy NE: (X, A) and (Z, D). We look now for a NE in mixed strategies of
the form

σ1 = pX + (1 − p)Z
σ2 = xA + (1 − x)D

The expected utilities of player 2 when he chooses pure actions are

u2 (σ1 , A) = 4p + (1 − p) = 1 + 3p
u2 (σ1 , D) = 0p + 3(1 − p) = 3 − 3p

We represent this three functions graphically,

11
 u2

3 u2 (σ1 , A) = 1 + 3p

u2 (σ1 , D) = 3 − 3p

3 1
3 1 p

We see that
0 ≤ p < 13

D
 (x = y = 0) if
BR2 (σ1 ) = {A, D} (x ∈ [0, 1]y = 0) if p = 31
1
A (x = 1) if 3 <p≤1


The expected utilities of player 1 when he chooses pure actions and y = 0 are

u1 (X, σ2 ) = x + 2(1 − x) = 2 − x
u1 (Z, σ2 ) = 0x + 3(1 − x) = 3 − 3x

Hence,
0 ≤ x < 21

Z  (p = 0) if
BR1 (σ2 ) = {X, Z} (p ∈ [0, 1]) if x = 21
X (p = 1) if 12 < x ≤ 1


Graphically,

BR2 (p)
1

1 BR1 (x)
2

1 X
3 1

And we obtain three NE

(a) x = y = 0; p = 0. That is, (Z, D), with payoffs u1 = u2 = 3.

12
 (b) x = 1, y = 0; p = 1. That is, (A, X), with payoffs u1 = 1, u2 = 4.
(c) x = 13 , y = 0; p = 21 . That is,
1 2
σ1 = X+ Z
3 3
1 1
σ2 = A+ D
2 2
with payoffs u1 = 23 , u2 = 2.

Problem 3: For the normal form game

X Y Z
A 10, 30 0, 20 20, 30
B 15, 35 10, 40 10, 40
C 25, 25 5, 25 5, 25

determine all the Nash equilibria of the game.

Solution: There are no (strictly) dominated strategies. The best replies of the players are

X Y Z
A 10, 30 0, 20 20, 30
B 15, 35 10, 40 10, 40
C 25, 25 5, 25 5, 25

We obtain three NE in pure strategies, (A, Z), (B, Y )and (C, X). We look now for a NE in mixed
strategies of the form

σ1 = xA + yB + (1 − x − y)C
σ2 = tX + sY + (1 − t − s)Z

The expected utilities of player 1 when he chooses pure actions are

u1 (A, σ2 ) = 10t + 20(1 − t − s) = 20 − 10t − 20s

u1 (B, σ2 ) = 15t + 10s + 10(1 − t − s) = 10 + 5t
u1 (C, σ2 ) = 25t + 5s + 5(1 − t − s) = 5 + 20t

The expected utilities of player 2 when he chooses pure actions are

u2 (σ1 , X) = 30x + 35y + 25(1 − x − y) = 25 + 5x + 10y

u2 (σ1 , Y ) = 20x + 40y + 25(1 − x − y) = 25 − 5x + 15y
u2 (σ1 , Z) = 30x + 40y + 25(1 − x − y) = 25 + 5x + 15y

Thus, the expected payoffs of the players under these strategies are,

u1 (σ1 , σ2 ) = 5 + 20t + 15x − 20sx − 30tx + 5y − 15ty

u2 (σ1 , σ2 ) = 25 + 5x − 10sx + 15y − 5ty

Method 1. We must have that both players are indifferent between the strategies used with (strictly)
positive probability.

Case 1: Suppose, there is a NE in mixed strategies in which player 1 is indifferent between strategies A,
B and C. We must have that
20 − 10t − 20s = 10 + 5t = 5 + 20t

13
So, t = 31 , s = 14 with expected utility u1 (A, σ2 ) = u1 (B, σ2 ) = u1 (C, σ2 ) = 35
3 . In this NE player 2 is
indifferent between strategies X, Y and Z. And we must have that

25 + 5x + 10y = 25 − 5x + 15y = 25 + 5x + 15y

But, this implies that x = y = 0. Hence, we conclude that there is no NE in which player 1 is indifferent
between strategies A, B and C. And we find the following NE

σ1 = C
1 1 5
σ2 = X+ Y + Z
3 4 12
with payoffs u1 = 35 3 , u2 = 25. Note for those values of t and s for which u1 (C, σ2 ) ≥ u1 (A, σ2 ) and
u1 (C, σ2 ) ≥ u1 (B, σ2 ) we also have that

σ1 = C
σ2 = tX + sY + (1 − t − s)Z

constitute a NE. That is, for those values of t, s ∈ [0, 1] such that t + s ≤ 1 and

20 − 10t − 20s ≤ 5 + 20t, 10 + 5t ≤ 5 + 20t,

the strategies σ1 and σ2 above constitute a NE. The above inequalities are equivalent to
1
t≥ , 3 ≤ 6t + 4s
3
and these are solved below in Case 1 of method 2. We obtain the following NE.

s1 = C
1 1 3 3
σ2 = tX + sY + (1 − t − s)Z, ≤t≤ , − t≤s≤1−t
3 2 4 2
and

s1 = C
1
σ2 = tX + sY + (1 − t − s)Z, ≤ t ≤ 1, 0≤s≤1−t
2
with payoffs u1 = 5 + 20t and u2 = 25.

Note that there are infinitely many NE of the form

Case 2:

s1 = B
1
σ2 = sY + (1 − s)Z, <s≤1
2
with payoffs u1 = 10, u2 = 40.

Case 3:

s1 = A
1
σ2 = tX + (1 − t)Z, 0≤t≤
2
with payoffs u1 = 20 − 10t, u2 = 30.

14
Case 4: Let us look for a NE in mixed strategies in which player 1 uses the two strategies A and C
with strictly positive probability. That is y = 0, 0 < x < 1. In this equilibrium, the expected utilities of
player 2 when he chooses pure actions are

u2 (σ1 , X) = 25 + 5x
u2 (σ1 , Y ) = 25 − 5x
u2 (σ1 , Z) = 25 + 5x

Since, u2 (σ1 , Y ) < u2 (σ1 , X) = u2 (σ1 , Z) for 0 < x < 1, the only possibility is that player 2 is using
strategies X and Z. That is, s = 0. The expected utilities of player 1 when he chooses pure actions are

u1 (A, σ2 ) = 20 − 10t
u1 (C, σ2 ) = 5 + 20t

And we must have 20 − 10t = 5 + 20t, That is t = 21 . That is, there are infinitely many NE of the form

σ1 = xA + (1 − x)C, 0≤x≤1
1 1
σ2 = X+ Z
2 2
The (expected) utilities of the players are

u1 (σ1 , σ2 ) = 15, u2 (σ1 , σ2 ) = 25 + 5x

Case 5: Let us look for a NE in mixed strategies in which player 1 uses the two strategies A and B with
strictly positive probability. That is y = 1 − x, 0 < x < 1. In this equilibrium, the expected utilities of
player 2 when he chooses pure actions are

u2 (σ1 , X) = 35 − 5x
u2 (σ1 , Y ) = 40 − 20x
u2 (σ1 , Z) = 40 − 10x

Note that u2 (σ1 , Y ) < u2 (σ1 , Z) for 0 < x < 1. Hence, player 2 will not use strategy Y . For player 2 to
be indifferent between X and Y we must have that 35 − 5x = 40 − 10x. That is, x = 1. Hence, there is no
NE in mixed strategies in which player 1 uses the two strategies A and B with strictly positive probability.

Let us look for a NE in mixed strategies in which player 1 uses the two strategies B and C with strictly
positive probability. That is x = 0, 0 < y < 1. In this equilibrium, the expected utilities of player 2 when
he chooses pure actions are

u2 (σ1 , X) = 25 + 10y
u2 (σ1 , Y ) = 25 + 15y
u2 (σ1 , Z) = 25 + 15y

Since, u2 (σ1 , X) < u2 (σ1 , Y ) = u2 (σ1 , Z) for 0 < y ≤ 1, the only possibility is that player 2 is using
strategies Y and Z. That is, t = 0. The expected utilities of player 1 when he chooses pure actions are

u1 (B, σ2 ) = 10
u1 (C, σ2 ) = 5

Since, u1 (B, σ2 ) > u1 (C, σ2 ), it is not possible that player 1 uses the strategies B and C. We conclude
that there is no NE in mixed strategies in which player 1 uses the two strategies B and C with strictly
positive probability.

15
Method 2. We use that (σ1∗ , σ2∗ ) is a NE iff σ1∗ ∈ BR1 (σ2 ) and σ2 ∈ BR2 (σ1∗ ). We use the notation
BR1 (s, t) = BR1 (σ2 ), BR2 (x, y) = BR2 (σ1 ). Player 1 chooses x, y ∈ [0, 1] with x + y ≤ 1. Player 2
chooses s, t ∈ [0, 1] with t + s ≤ 1. Recall that

u1 (x, y; s, t) = 5 + 20t + 15x − 20sx − 30tx + 5y − 15ty

u2 (x, y; s, t) = 25 + 5x + 15y − 5ty − 10sx

Case 1: Let us look for a NE in which x = y = 0. The utility of player 2 is

u2 (0, 0; s, t) = 25

And player 2 is indifferent among any of his strategies. The utility of player 1 may be written as

u1 (x, y; s, t) = 5 + 20t + 15x − 20sx − 30tx + 5y − 15ty

= 5 + 20t + 5x(3 − 4s − 6t) + 5y(1 − 3t)

The action x = y = 0 is a best response for player 1 only if 3 − 4s − 6t ≤ 0 and 1 − 3t ≤ 0. Graphically,

3/4

s+t=1

3 − 4s − 6t = 0

1/3 1/2 1 t

The green area represents the values of s, t for which we have 3 − 4s − 6t ≤ 0 and 1 − 3t ≤ 0. This region
may be described as
1 1 3 3 1
A = {(s, t) ∈ R2 : ≤t≤ , − t ≤ s ≤ 1 − t} ∪ {(s, t) ∈ R2 : ≤ t ≤ 1, 0 ≤ s ≤ 1 − t}
3 2 4 2 2
Thus, we find infinitely many NE of the form

s1 = C
σ2 = tX + sY + (1 − t − s)Z, t, s ∈ A

The utilities of the players are

u1 = 5 + 20t u2 = 25

Let us look for a NE in which x = 0, y > 0. The utility of player 2 is

u2 (0, y; s, t) = 25 + 15y − 5ty

Since, y > 0, player 2 chooses t = 0. The utility of the player 1 is

u1 (x, y; s, 0) = 5 + 15x − 20sx + 5y = 5 + x(15 − 20s) + 5y

16
We see that if s ≥ 34 , then player 1 chooses x = 0, y = 1. If s < 3
4 , then we may assume x + y = 1 and

u1 (x, 1 − x; s, 0) = 10 + x(10 − 20s)

and we see that player 1 chooses

• x = 0, y = 1, if s > 21 ;
• x = 1, y = 0, if s < 12 ;
• x ∈ [0, 1], y = 1 − x, if s = 12 .

That is, 
1
x = 0, y = 1
 if s> 2
1
BR1 (t = 0, s) = x = 1, y = 0 if s< 2
1

x ∈ [0, 1], y = 1 − x if s=

2

Case 2: Now we check wether

1
x = 0, y = 1, t = 0, s ≥
2
is NE. If player 1 chooses x = 0, y = 1, the utility of player 2 is

u2 (0, 1; s, t) = 40 − 5t
1
so his best response is to choose t = 0 and any s. If Player 2 chooses t = 0, s ≥ 2 we have already seen
that player 1 chooses y = 1. Hence, we have infinitely many NE of the form

s1 = B
1
σ2 = sY + (1 − s)Z, ≤s≤1
2
The utilities of the players are
u1 = 10 u2 = 40

Let us look for a NE in which x > 0, y = 0. The utility of player 2 is

u2 (x, 0; s, t) = 25 + 5x − 10sx

Since, x > 0, player 2 chooses s = 0. The utility of the player 1 is

u1 (x, y; 0, t) = 5 + 20t + 15x − 30tx + 5y − 15ty = 5 + 20t + x(15 − 30t) + 5y(1 − 3t)

Note that,

1
(i) For t > 2 the coefficients of x and y are negative and player 1 chooses x = y = 0.
1
(ii) If t = 2 we have  
1 5y
u1 x, y; , 0 = 15 −
2 2
so player 1 chooses y = 0, any x ∈ [0, 1].
(iii) For 13 < t < 12 the coefficient of y is negative and the coefficient of x is positive. Hence, player 1
chooses x = 1, y = 0.
1
(iv) For t = 3 the coefficient of y is 0 and the coefficient of x is positive. Hence, player 1 chooses x = 1,
y = 0.
1
(v) For t < 3 the coefficients of y and x are both positive. We may assume y = 1 − x and

u1 (x, 1 − x; t, 0) = 10 + 5t + 10x − 15tx = 10 + 5t + x(10 − 15t)

Hence, player 1 chooses x = 1, y = 0.

17
That is, 
x = y = 0
 if t > 21
BR1 (t, s = 0) = y = 0, x ∈ [0, 1] if t = 21
1

x=1 if 0≤t<

2

Case 3: Now we check wether

1
x = 1, y = 0, s = 0, t ≤
2
is a NE. If player 1 chooses x = 1, y = 0, the utility of player 2 is

u2 (1, 0; s, t) = 30 − 10s

So, player 2 chooses s = 0 and any t ∈ [0, 1]. We have infinitely many NE of the form

s1 = A
1
σ2 = tX + (1 − t)Z, 0≤t≤
2
The utilities of the players are
u1 = 20 − 10t u2 = 30

Case 4: We check if there is a NE with

1
y = 0, x ∈ (0, 1], s = 0, t =
2
The utility of player 1 is  
1 5y
u1 x, y; 0, = 15 −
2 2
So, player 1 chooses y = 0 and any x ∈ [0, 1]. For y = 0 the utility of player 2 is

u2 (x, 0; s, t) = 25 + 5x − 10sx

so player 2 chooses s = 0 and any t ∈ [0, 1]. Therfore, there are infinitely many NE of the form

σ1 = xA + (1 − x)C, 0≤x≤1
1 1
σ2 = X+ Z
2 2
The (expected) utilities of the players are

u1 (σ1 , σ2 ) = 15, u2 (σ1 , σ2 ) = 25 + 5x

Case 5: Let us look for a NE in which x > 0, y > 0. In this case, player 2 chooses t = s = 0. The
utility of player 1 is
u1 (x, y; 0, 0) = 5 + 15x + 5y
we may assume y = 1 − x ,
u1 (x, 1 − x; 0, 0) = 10 + 10x
and player 1 chooses x = 1, y = 0. There is no such NE.
Problem 4: Consider a game in which there is a prize worth $30. There are three contestants, A, B,
and C. Each can buy a ticket worth $15 or $30, or not buy a ticket at all. They make these choices
simultaneously and independently. Once the ticket purchase decisions have been made, the prize is
awarded. If no one bought the ticket, the prize is not awarded. Otherwise, the ticket is awarded to the
buyer of the higher cost ticket if there is only one such player, or split equally between two or three
players if there are ties among the highest cost ticket buyers.

18
 1. Write the normal form of the game. (Hint: Draw a box for each of player C?s possible strategies.)
2. Identify the Nash equilibria in pure strategies of this game.

Solution: Player C chooses one of the following 3 matrices

Player B
0 15 30
(a) Player C chooses 0: 0 0, 0, 0 0, 15, 0 0, 0, 0
Player A 15 15, 0, 0 0, 0, 0 −15, 0, 0
30 0, 0, 0 0, −15, 0 −15, −15, 0

Player B
0 15 30
(b) Player C chooses 15: 0 0, 0, 15 0, 0, 0 0, 0, −15
Player A 15 0, 0, 0 −5, −5, −5 −15, 0, −15
30 0, 0, −15 0, −15, −15 −15, −15, −15

Player B
0 15 30
(c) Player C chooses 30: 0 0, 0, 0 0, −15, 0 0, −15, −15
Player A 15 −15, 0, 0 −15, −15, 0 −15, −15, −15
30 −15, 0, −15 −15, −15, −15 −20, −20, −20

The best responses of the players are

Player B
0 15 30
(a) Player C chooses 0: 0 0, 0, 0 0, 15, 0 0, 0, 0
Player A 15 15, 0, 0 0, 0, 0 −15, 0, 0
30 0, 0, 0 0, −15, 0 −15, −15, 0

Player B
0 15 30
(b) Player C chooses 15: 0 0, 0, 15 0, 0, 0 0, 0, −15
Player A 15 0, 0, 0 −5, −5, −5 −15, 0, −15
30 0, 0, −15 0, −15, −15 −15, −15, −15

Player B
0 15 30
(c) Player C chooses 30: 0 0, 0, 0 0, −15, 0 0, −15, −15
Player A 15 −15, 0, 0 −15, −15, 0 −15, −15, −15
30 −15, 0, −15 −15, −15, −15 −20, −20, −20

We see that there are 6 NE in pure strategies of to types:

Type 1: One player plays 0 and the other two play 15.
Type 2: One player plays 15 and the other two play 0.

Problem 5: Two firms, 1 and 2, produce heterogeneous products. If the two firms set prices p1 and p2 ,
respectively, the quantities demanded of the products of the two firms will be,
  
p1 + p2
x1 (p1 , p2 ) = max 0, 180 − p1 − p1 −
2
  
p1 + p2
x2 (p1 , p2 ) = max 0, 180 − p2 − p2 −
2

19
The above demand functions describe a situation in which the products are not perfectly homogeneous.
For instance, the demand for good 1, even if decreasing in the price set by firm 2, is positive also when
p1 > p2 (which is not the case when the good are homogeneous), or, in a different way, the demand for
good 1 will be lower than 180 − p1 when the price of good 1 is higher than the mean of the prices (and
therefore higher than p2 ) and higher than that when the price of good 1 is lower than the mean of the
prices (and therefore lower than p2 ). Assume that the firms have constant marginal costs MC1 = 20
and MC2 = 20. Assume also that the characteristics of the market are such that the firms have to set
simultaneously the prices.

(a) Write down firms’ profits, π1 (p1 , p2 ) and π2 (p1 , p2 ), as functions of the prices they set.
(b) Find firm 1’s best response to a price set by firm 2, BR1 (p2 ).
(c) Find firm 2’s best response to a price set by firm 1, BR2 (p1 ).

(d) Find the Nash equilibrium (p∗1 , p∗2 ).

(e) Determine equilibrium profits, π1 (p1 ,∗ p∗2 ) and π2 (p∗1 , p∗2 ).
(f) Are equilibrium profits zero? Why?
(g) Draw the best response function and the Nash equilibrium in the p2 − p1 plane (p2 on the horizontal
axis and p1 on the vertical axis).

Solution:

(a) We assume prices are such that the quantities demanded are positive. The profits of the firms are

3p2
  
p1 + p2 p1 p2
π1 (p1 , p2 ) = 180 − p1 − p1 − (p1 − 20) = − 1 + + 210p1 − 10p2 − 3600
2 2 2
3p2
  
p1 + p2 p1 p2
π2 (p1 , p2 ) = 180 − p2 − p2 − (p2 − 20) = − 10p1 − 2 + 210p2 − 3600
2 2 2

(b) Firm 1’s best response to a price set by firm 2 is

p2 + 420
BR1 (p2 ) =
6

(c) Fiirm 2’s best response to a price set by firm 1 is

p1 + 420
BR2 (p1 ) =
6

(d) The Nash equilibrium (p∗1 , p∗2 ) is the solution of the following system of equations
p2 + 420
p1 =
6
p1 + 420
p2 =
6
The solution is p∗1 = p∗2 = 84.
(e) The equilibrium profits are π1 (p∗1 , p∗2 ) = π2 (p∗1 , p∗2 ) = 6144.
(f ) The equilibrium profits aren’t zero because products are heterogeneous.

(g) Graphically, the best response function and the Nash equilibrium are

20
 p1

BR2 (p1 )
84
BR1 (p2 )
70

70 84 p2

Problem 6: Two firms, 1 and 2, produce heterogeneous products. If the two firms set prices p1 and p2 ,
respectively, the quantities demanded of the products of the two firms will be:
 
p1 + p2
x1 (p1 , p2 ) = 140 − p1 − p1 −
2
 
p1 + p2
x2 (p1 , p2 ) = 140 − p2 − p2 −
2

Suppose the two firms have constant marginal costs MC1 = 40 and MC2 = 40 and that they have to set
their prices simultaneously.

(a) What is the Nash equilibrium ? What are the Nash equilibrium profits for firms 1 and 2?
(b) Are the profits in equilibrium equal to 0? Why or why not?

Solution:

(a) We assume prices are such that the quantities demanded are positive. The profits of the firms are

3p2
  
p1 + p2 p1 p2
π1 (p1 , p2 ) = 140 − p1 − p1 − (p1 − 40) = − 1 + + 200p1 − 20p2 − 5600
2 2 2
3p2
  
p1 + p2 p1 p2
π2 (p1 , p2 ) = 140 − p2 − p2 − (p2 − 40) = − 20p1 − 2 + 200p2 − 5600
2 2 2

Firm 1’s best response to a price set by firm 2 is

p2 + 420
BR1 (p2 ) =
6

Fiirm 2’s best response to a price set by firm 1 is

p1 + 400
BR2 (p1 ) =
6

The Nash equilibrium (p∗1 , p∗2 ) is the solution of the following system of equations
p2 + 420
p1 = .
6
p1 + 400
p2 =
6
The solution is p∗1 = p∗2 = 80. The equilibrium profits are π1 (p∗1 , p∗2 ) = π2 (p∗1 , p∗2 ) = 2400.

21
(b) The equilibrium profits aren’t zero because products are heterogeneous.

Problem 7: Assume that two firms, 1 and 2, produce heterogeneous products and the quantities de-
manded by the market, when these firms fix prices p1 and p2 , are, respectively:

xi1 (p1 , p2 ) = 200 − p1 + p2 /3

x2 (p1 , p2 ) = 200 − p2 + p1 /3

These demand functions describe a situation in which products are not perfectly homogenous. Suppose
that both firms have constant marginal costs MC= 50 and MC2 = 50 and that the features of the market
are such that both firms have to set prices simultaneously.

(a) Find the Nash equilibrium (p∗1 , p∗2 ).

(b) Determine equilibrium profits, π1 (p1 ,∗ p∗2 ) and π2 (p∗1 , p∗2 ). Are equilibrium profits zero? Why?
(c) Suppose now that firm 1 and 2 could make an agreement on how to make their products different so
that the demand functions become,

xi1 (p1 , p2 ) = 200 − p1 + p2 /2

x2 (p1 , p2 ) = 200 − p2 + p1 /2

Do you think that it pays for the firms to agree on these new differentiated products?

Solution: Let us consider the following market demand functions

x( p1 , p2 ) = 200 − p1 + bp2
x2 (p1 , p2 ) = 200 − p2 + bp1

with 0 ≤ b < 2. We assume prices are such that the quantities demanded are positive. Now, the profits
of the firms are

π1 (p1 , p2 ) = (200 − p1 + bp2 ) (p1 − 50) = bp1 p2 − 50bp2 − p21 + 250p1 − 10000
π2 (p1 , p2 ) = (200 − p2 + bp1 ) (p2 − 50) = bp1 p2 − 50bp1 − p22 + 250p2 − 10000

Firm 1’s best response to a price set by firm 2 is

1
BR1 (p2 ) = (bp2 + 250)
2

Fiirm 2’s best response to a price set by firm 1 is

1
BR2 (p1 ) = (bp1 + 250)
2

The Nash equilibrium (p∗1 , p∗2 ) is the solution of the following system of equations
1
p1 = (bp2 + 250)
2
1
p2 = (bp1 + 250)
2
The solution is
250
p∗1 = p∗2 =
2−b

The equilibrium profits are

2500(b + 3)2
π1 (p∗1 , p∗2 ) = π2 (p∗1 , p∗2 ) =
(2 − b)2

22
Note that the funcion
2500(b + 3)2
f (b) =
(2 − b)2
is increasing in b. Its graph is the following.

1 × 106

800 000

600 000

400 000

200 000

0.5 1.0 1.5 2.0

(a) Taking
1
b=
3
we obtain that the Nash equilibrium (p1 , p2 ) is is p1 = p∗2 = 150.
∗ ∗ ∗

(b) Taking
1
b=
3
the equilibrium profits are π1 (p∗1 , p∗2 ) = π2 (p∗1 , p∗2 ) = 10000. The equilibrium profits aren’t zero because
products are heterogeneous.

(c) Suppose now that firm 1 and 2 could make an agreement on how to make their products different so
that the demand functions become,

x( p1 , p2 ) = 200 − p1 + p2 /2
x2 (p1 , p2 ) = 200 − p1 + p1 /2

Now, taking
1
b=
2
The Nash equilibrium is p∗1 = p∗2 = 460 ∗ ∗ ∗ ∗
3 . And the equilibrium profits are π1 (p1 , p2 ) = π2 (p1 , p2 ) =
96100
9 ≈ 10677.61. Hence, the firms would benefit from product differentiation.

Problem 8: Firms 1 and 2 compete in a market with a homogeneous product. Firms have to decide
how much to produce and put up for sale. Let q1 and q2 denote the quantities produced and put up for
sale by firm 1 and 2, respectively. The clearing market price of the homogeneous product on sale depends
on quantities produced according to the following inverse demand function: P (Q) = max{0, 10 − Q},
where Q = q1 + q2 is the total number of units of the product in the market. Total production costs are
c1 (q1 ) = q12 and c2 (q2 ) = q22 . This is the Cournot model of oligopoly competition that you will study in
the course of Industrial Organization.

(a) Describe the situation as a normal form game.

23
(b) Find the Nash equilibrium.

Solution:

(a) The set of players is N = {1, 2}, the sets of strategies are S1 = S2 = [0, ∞) and the payoffs are

π1 (q1 , q2 ) = (10 − q1 − q2 ) q1 − q12

π2 (q1 , q2 ) = (10 − q1 − q2 ) q2 − q22

(b) The profits of the firms are

π1 (q1 , q2 ) = (10 − q1 − q2 ) q1 − q12 = q1 (−q1 − q2 + 10) − q12

π2 (q1 , q2 ) = (10 − q1 − q2 ) q2 − q22 = q2 (−q1 − q2 + 10) − q22

Firm 1’s best response to a price set by firm 2 is

10 − q2
BR1 (q2 ) =
4

Firm 2’s best response to a price set by firm 1 is

10 − q1
BR2 (q1 ) =
4

The Nash equilibrium (q1∗ , q2∗ ) is the solution of the following system of equations
10 − q2
q1 =
4
10 − q1
q2 =
4
The solution is q1∗ = q2∗ = 2. The Equilibrium profits are π1∗ = π2∗ = 8.

24