Topic 2: Static Games of Complete Information

2.1 The Normal-form Representation

The normal form of a game is defined as 𝐺 = {𝑆 , 𝑆 , … , 𝑆 ; 𝑢 , 𝑢 , … , 𝑢 }.

Let's consider a classic example known as the Prisoner's Dilemma to illustrate the
normal-form representation. In this game, two individuals are arrested and charged
with a crime. They are placed in separate cells.

 The payoff is negative. For example, if a player receives a 10-year sentence, his
payoff is -10.
 If both confess to the crime: each receives a payoff of -10.
 If both remain silent: each receives a payoff of -3.
 If only one confess and the other one remains silent, the one who confesses
receive a payoff of -1 and the other one who remains silent receive a payoff of


-20. An action n Payoff
There is complete information: each player’s payoff function is common
knowledge among all the players.

Player 2
Silent (S) Confess (C)
Player 1 Silent (S) (-3, -3) (-20, -2)
Confess (C) (-1, -20) (-10, -10)

In this normal-form representation, the rows represent Player 1's strategies (S and C),
while the columns represent Player 2's strategies (S and C). The numbers in each cell
represent the payoffs for Player 1 and Player 2, respectively, based on their chosen
strategies.

To solve this game, we need an equilibrium concept.

1
O
a) si (silent , Confess
32 (silent, contesss

2.1.1 ICE
4)(3)() , ((( , 3)(4 , )

Referring to the above Prisoner's Dilemma:

~ sig sa , selb ,
2

a. Write down the set of strategies of each player.

b. Write down all the possible outcomes.
-
c. What are the payoffs in 𝑢 (𝑆, 𝐶), 𝑢 (𝐶, 𝐶), 𝑢 (𝑆, 𝑆), and 𝑢 (𝐶, 𝑆)?
UILS () ,
: -
20
Answers
= - 10available to player 1 (𝑆 ) is S and C. The set of strategies
UI (4
a. The set , 4
of strategies
available to player 2 (𝑆 ) is S and C.
U2(S
b. Possible3) = 3 -

outcomes of this games are: (𝑆, 𝑆), (𝑆, 𝐶), (𝐶, 𝑆), (𝐶, 𝐶).
,
c. 𝑢 (𝑆, 𝐶) = −20, 𝑢 (𝐶, 𝐶) = −10, 𝑢 (𝑆, 𝑆) = −3, and 𝑢 (𝐶, 𝑆) = −20.
U2(4 3) = ,
So -

2.2 Equilibrium Concept 1: Iterated Elimination of Strictly

Dominated Strategies &K

2.2.1 Strictly Dominated Strategies

A strictly dominated strategy is always worse than some other strategy regardless of
the strategies of the other players are. For example, choosing high price yields a lower
payoff regardless of whether the other company choose high price or low price. Then,
high price is a strictly dominated strategy.

2.2.2 Iterated Elimination of Strictly Dominated Strategies

eliminate e
Y
lint
Now if we postulate that rational players do not play strictly dominated strategies
then use this assumption to solve the Prisoners’ Dilemma.
HE MAY
EXcliente ente
Player
Player 2
1
Silent (S) Confess (C)
Player 1 Silent (S) ~
(-3, -3) -
(-20, -2)
Confess (C)
G
(-1, -20)
②
(-10, -10)

For player 1, silent is always worse than confess regardless of whether player 2 plays
vilient/c
Silent : dominate
2
play2
I EquiE & PFK
player 3
 1 : -1323
- **
player -103 -
20 Silent -
10
-
10.

silent or confess. We can eliminate silent.

↓
EjF Player 2
Silent (S) Confess (C)
Player 1 Silent (S) (-3, -3) (-20, -2)
Confess (C)
O
(-1, -20)
&
(-10, -10)

For player 2, silent is always worse than confess. We can eliminate silent. The
equilibrium is 𝒔∗ = (𝑪, 𝑪). 2
player
-10) 20 -

Player 2
Silent
#J
Wor
Silent (S) Confess (C)
Player 1 Silent (S) (-3, -3) (-20, -2)
Confess (C) (-1, -20) C
(-10, -10)

2.2.2.1 ICE
/

Solve the game below using by the iterated elimination of strictly dominated strategies.
D Player 2 ④

I
Left Middle Right

01
Player 1 Up (1, 0) (1, 2) (2, 1)
-
Down -
(0, 3) -
(0, 1) -
(0, 0)

2.3 Equilibrium Concept 2: The Nash Equilibrium

best response ,

However, IESDS does not always solve the game. For example, consider the following:
It strindom
m
for Player I
Player 2
Left Middle Right
Player 1 Up (1, 0) (1, 2) (0,
* 1)
Down (0, 3)
p (0,
def 1) (2, 2)

When IESDS fails we need to strengthen the equilibrium concept further in order to
solve the second example. This leads to our next equilibrium concept: the Nash
Equilibrium.
3
2.3.1 Best Response Analysis

The strategies of a game are a Nash Equilibrium if, for each player, the equilibrium
strategy is the best (utility or payoff maximizing) response to the equilibrium strategy
of all other players. In other words, at a Nash equilibrium, players play their mutual
best responses: “my action is a best response to your action” and “your action is a
best response to my action.” -
F* 7
OFF plae
Now, instead of finding the worst strategy, we find the best strategy regardless of
the strategies of the other players are.

Price war: Consider a one-shot simultaneous-move game with complete information.

Company A (row player) and Company B (column player), making pricing decisions for
their products. Below is the normal-form representation of the game.

FT , *B FR Player
, 2
8>+ Company B 3/
High (H) Low (L)
Company A High (H) (5, 5) (1,↳
8) 8)
Low (L)
8
(8, 1) OG
(3, 3) 3
Knash
=

For Company A: -

 If Company B chooses H, Company A's best response is to choose L since 185

> 5.
 If Company B chooses L, Company A's best response is to choose L since 3·
> 8.

For Company B:
 If Company A chooses H, Company B's best response is to choose L since 8 > 5.
 If Company A chooses L, Company B's best response is to choose L since 3 > 1.

The Nash equilibrium occurs when both companies choose their best responses to
more
each other's strategies. In this case, the only outcome that satisfies this condition is
when both companies choose the strategy L. Therefore, (L, L) is the pure strategy Nash
equilibrium. A pure strategy refers to a player chooses a specific action with certainty.
-

-
2.3.1.1 Real world business example of Prisoner's Dilemma: price war

The two biggest IT companies in China, Alibaba and Tencent once fought against each
other in the market of taxi app in China. Both of them captured about half of the
market share before the war and afterwards.

4
They encouraged people to use their apps and thus to stay on their app’s platform by
subsiding both users and taxi-drivers. At first the amount of subsidy for each
transaction was small. Nonetheless, the amount of subsidy kept growing for about 6
months. It even made a considerable amount of rides for passengers free, meanwhile
the taxi-drivers also earned more than the taxi price.

Finally, both of them called off this insane war almost at the same time, and reduced
the subsidies to a normal level, leaving hundreds of thousands of customers
complaining about the reduction, while they still have to use one of the two.

2.3.2 Multiple Nash Equilibria

In the above example, there is only one Nash equilibrium. Now, consider the following
battle of the sexes. A couple Pat (column player) and Chris (row player) have the
following preferences for a night’s entertainment: Both prefer to be together than
apart irrespective of the choice of outing. Pat prefers opera to football, and Chris
prefers football to opera. These preferences can be presented as payoffs and the game
in a normal form can be represented as follows:
Pat
Fl nash
Opera (O) Football (F)
Chris Opera (O) 08
(1, 2) (0, 0)
Football (F) (0, 0)
KO
(2, 1)

In this case, there are two pure Nash equilibria. Can you find them?

-
2.3.3 Business example of the battle of the sexes

One common example in business is industry standards. Suppose company A and B

own the two major messaging platforms in some region. Company A has already
invested heavily in developing and implementing Protocol X, while Company B have
invested in Protocol Y. However, they realize that having incompatible protocols
creates friction for users and hinders widespread adoption of their services. it is costly
for the companies to transition their existing infrastructure to a new protocol, but the
cost of doing so is smaller than the benefits of coordinating on a common
communication protocol.

5
 don't
2.3.4 Mixed Strategies player I
want Yo
↑

match
Some games have no pure Nash equilibrium. One example is known as Matching
Pennies. Each player’s strategy set is {Heads, Tails}. Each player displays her choice of
Head or Tail. If the two pennies match, Player 1 wins Player 2’s penny, if they do not
match Player 2 wins Player 1’s penny.

no

↓
The payoff matrix is as follows:

b Pr
of
Player 2
Heads (H) Tails (T)
equ
Player 1 tHeads (H)
8(1, -1) (-1, 1)
(-1,O
Tails (T) 1)
G
(1, -1)

For Player 1:
 If Player 2 chooses Heads, Player 1's best response is to choose Heads (1 > -1).
 If Player 2 chooses Tails, Player 1's best response is to choose Tails (1 > -1).

For Player 2:
 If Player 1 chooses Heads, Player 2's best response is to choose Tails (1 > -1).
 If Player 1 chooses Tails, Player 2's best response is to choose Heads (1 > -1).

-
There is no pure Nash equilibrium because the best response for each player depends
on the other player's choice. A pure strategy refers to a player chooses a specific action
with certainty.
* iijn FiR15) T option -
-

In Matching Pennies, we play the game in the following way: When Player 1 chooses
H, Player 2 must choose T. Player 1 chooses T, Player 2 must choose H. Each player
chooses a specific action with certainty. We find that there is no (pure Nash)
equilibrium.

But how about if we add uncertainty to the players’ strategies? When Player 1 chooses
H, Player 2 will choose T with a probability and choose H with another probability.
When Player 1 chooses T, Player 2 will choose T with a probability and choose H with

Tayers
another probability. In this case, Player 2 is “mixing the strategies”. Of course, Player
1 can also do the same (mixing the strategies). Can this result an equilibrium?

ertian
Komverthing
6
2.3.4.1 The Mixed Strategies Nash Equilibrium for Matching Pennies
Fi
Now, we consider mixed strategies for this game. This means we allow players to
choose each action with a probability.

For Player 1:
 Let 𝒑 be the probability of choosing H. Then 𝟏 − 𝒑 is the probability of
choosing T.

For Player 2:
 Let 𝒒 be the probability of choosing H. Then 𝟏 − 𝒒 is the probability of
choosing T.

Player 2
𝒒 𝟏−𝒒
7
Heads (H) Tails (T)↑
Player 1 𝒑 Heads (H) (1, -1) (-1, 1)
𝟏−𝒑 E Tails (T) (-1, 1) (1, -1)

The game is now played under uncertainty. According to the rationality assumption,
when there is uncertainty, rational players maximize expected utility or payoff.

7
 Player 2
𝒒 𝟏−𝒒

-X
Heads (H) Tails (T)
Player 1 𝒑 Heads (H) (1, -1) (-1, 1)
𝟏−𝒑 Tails (T) (-1, 1) (1, -1)

O
Consider Player 1. Compare expected payoffs from H and T.

Expected payoffs from H:

FT ***
1 × 𝑞 + (−1) × (1 − 𝑞) = 2𝑞 − 1
Expected payoffs from T:
−1 × 𝑞 + 1 × (1 − 𝑞) = 1 − 2𝑞

If the expected payoffs from H > the expected payoffs from T, Player 1 will choose H.
1
2𝑞 − 1 > 1 − 2𝑞 ⟹ 𝑞 >
2

-
When 𝑞 > , Player 1 must choose H.

-
If the expected payoffs from H < the expected payoffs from T, Player 1 will choose T.
1
2𝑞 − 1 < 1 − 2𝑞 ⟹ 𝑞 <
2

When 𝑞 < , Player 1 must choose T.

-

If the expected payoffs from H = the expected payoffs from T, Player 1 is indifferent

It
between choosing H or T.
1
2𝑞 − 1 = 1 − 2𝑞 ⟹ 𝑞 =
2

When 𝑞 = , Player 1 is indifferent between choosing H or T.

H
H/T

8
 Player 2
𝒒 𝟏−𝒒

Player 1 𝒑
-
Heads (H) (1, O
-1)
Heads (H) Tails (T)
(-1, 1)
𝟏−𝒑 Tails (T) O
(-1, 1) (1, -1)
-
Consider Player 2. Compare expected payoffs from H and T.

Expected payoffs from H:

−1 × 𝑝 + 1 × (1 − 𝑝) = 1 − 2𝑝
player
Expected payoffs from T:
1 × 𝑝 + (−1) × (1 − 𝑝) = 2𝑝 − 1

If the expected payoffs from H > the expected payoffs from T, Player 2 will choose H.
1
1 − 2𝑝 > 2𝑝 − 1 ⟹ 𝑝 <
2
-
When 𝑝 < , Player 2 must choose H.

If the expected payoffs from H < the expected payoffs from T, Player 2 will choose T.
1
1 − 2𝑞 < 2𝑞 − 1 ⟹ 𝑝 >
2

When 𝑝 > , Player 2 must choose T.

If the expected payoffs from H = the expected payoffs from T, Player 2 is indifferent
between choosing H or T.

1
1 − 2𝑞 = 2𝑞 − 1 ⟹ 𝑝 =
2

When 𝑝 = , Player 2 is indifferent between choosing H or T.

9
The below is a summary. # p15) P -

1.
𝑞>
1
2
①
Player 1 must choose H

2. 1 Player 1 must choose T

FE?!
𝑞<
2
3. 1 Player 2 must choose H
𝑝>
2

Pet
4. 1 Player 2 must choose T
𝑝<
2
5. O1 Player 1 is indifferent between H and T
𝑞=
2
6.
② 𝑝=
1
2
Player 2 is indifferent between H and T

Clearly, 1 – 4 are not mixed strategies Nash equilibria. Why? For example, when 𝒒 >
-
𝟏
, Player 1 must choose H. Player 1 doesn’t “choose each action with a probability.”
𝟐

-
In 5 and 6, when each player is indifferent between choosing H and T, they will not
choose H or T with certainty. In other words, they will choose each action with a
probability (mixed strategies). E probE FREPI
action .

The mixed strategies Nash equilibrium is Player 1 chooses H with a probability of 0.5
and T with a probability of 0.5, while Player 2 chooses H with a probability of 0.5 and
T with a probability of 0.5. The formal presentation of a mixed strategies Nash

probability
equilibrium is:
-

1 1 1 1
(𝑝, 1 − 𝑝), (𝑞, 1 − 𝑞) = , , ,
2 2 2 2

10
2.3.4.2 Procedure for Solving the Mixed Strategies Nash Equilibrium

After understanding the logic of the mixed strategies Nash equilibrium, the math
become relatively straightforward.

Player 2
𝒒 𝟏−𝒒
Heads (H) Tails (T)
Player 1 𝒑 Heads (H) (1, -1) (-1, 1)
𝟏−𝒑 Tails (T) (-1, 1) (1, -1)

Using Player 1 as an example:

8
Step 1: Write down the expected payoffs from each action (H and T) for each player.

Expected payoffs from H: player I

s
1 × 𝑞 + (−1) × (1 − 𝑞) = 2𝑞 − 1
Expected payoffs from T:
−1 × 𝑞 + 1 × (1 − 𝑞) = 1 − 2𝑞

Step 2: Apply the indifference condition. -E

1
2𝑞 − 1 = 1 − 2𝑞 ⟹ 𝑞 =
2

When 𝑞 = , Player 1 is indifference between choosing H and T and he will play a

mixed strategy.

·
Step 3: Repeat Step 1 and 2 for Player 2.

The mixed strategies Nash equilibrium is (𝑝, 1 − 𝑝), (𝑞, 1 − 𝑞) = , , , .

Note that it is possible for a game to have both pure Nash equilibrium and mixed
-
strategies Nash equilibrium. The below exercise is an example.
-

11
 men
9955
+ 2.3.4.3 ICE
g
-
Consider again the battle of the sexes.
1 4
- -

4 Pat
Opera (O) Football (F)

p
Chris Opera (O) (1, 2) (0, 0)
Football (F) (0, 0) (2, 1)
For Chris:
 Let 𝑝 be the probability of choosing O. Then 1 − 𝑝 is the probability of
choosing F.
For Pat:
 Let 𝑞 be the probability of choosing O. Then 1 − 𝑞 is the probability of
choosing F.

Find the mixed strategies Nash equilibrium.

Chris
Answers

expected 1 4 (0) x (1 4)
For Chris:
x +
-

;
Expected payoffs from O:
1 × 𝑞 + (0) × (1 − 𝑞) = 𝑞
&
from Expected payoffs from F:
0 × 𝑞 + 2 × (1 − 𝑞) = 2 − 2𝑞

F Apply
0x4
: the indifference
22
+ condition: 1 - p32224 2

=
𝑞 = 2 − 2𝑞 ⟹ 𝑞 =
: 3
422-24
When 𝑞 = , Playerp1 is indifference between choosing O and F and he will play a

when 42t
mixed strategy. ,
play 1 is indiff
For Pat:
Pat O:
2p 01O: 1
+ from
Expected payoffs
-

P) =
<P
2 × 𝑝 + (0) × (1 − 𝑝) = 2𝑝

F 0(p)
Expected payoffs from F:
: +
0 × 𝑝 +p)
1(/ - =
IvP
1 × (1 − 𝑝) = 1 − 𝑝

Pr5
Apply the indifference condition:
24 = / -

P 2𝑝 = 1 − 𝑝 ⟹ 𝑝 =
1

(5 3) It 5
3

4) (4612q)
1 mixed strategies :
<P ,
The
-
, Nash equilibrium is (𝑝, 1 − 𝑝), (𝑞,
,
1 − 𝑞) = , ,, ,

12
2.3.5 No Profitable Unilateral Deviation

At the Nash equilibrium, players have no incentive to deviate unilaterally. That is, if
all other players are playing their Nash equilibrium actions, I cannot do strictly better
by unilaterally changing my action. For example, in the Prisoner's Dilemma, the pure
Nash equilibrium is (C, C).

Player 2
Silent (S) Confess (C)
Player 1 Silent (S) (-3, -3) (-20, -2)
Confess (C) (-1, -20) (-10, -10)
-

Let’s check whether there is a profitable unilateral deviation. Suppose Player 1

unilaterally deviates to play S while Player 1 remains to play C. Then, Player 2’s payoff
changes from -10 to -20. Player 2 is not better off and there is no profitable unilateral
deviation. Same happens for Player 1. Therefore, (C, C) is a pure Nash equilibrium.

We can use this property in order to locate the Nash equilibrium. Instead of directly
finding the best responses (that we have done so far) in finding the Nash equilibrium,
we can start with any action combination and ask whether it is a Nash equilibrium or
not by checking whether there exists any incentive for at least one of the players to
deviate unilaterally and obtain a higher payoff.

Back to the Battle of Sexes game. The following presents the normal-form
representation of the game.

-
Pat 0 2

Al
58
Opera (O) Football (F)
,

O
GFootball (F) O
Chris Opera (O) (1, 2) (0, 0)
(0, 0) (2, 1)

Let’s consider the action choices (O, O) in the battle of the sexes. Given that Pat
chooses Opera (O), if Chris unilaterally changes his action to Football (F), he will
receive 0 rather than 1. Thus, there is no profitable unilateral deviation for Chris.

Similarly, given that Chirs chooses Opera (O), if Pat unilaterally changes her action to
Football (F), she will receive 0 rather than 2. Thus, there is no profitable unilateral
deviation for Pat. (O, O) is a pure Nash equilibrium.
13
 2.3.5.1 ICE

Let’s consider the action choices (F, O) and (F, F) in the battle of the sexes. Show that
whether they are a pure Nash equilibrium by using the method of no profitable
unilateral deviation.

Answers(F 0)
For

patchooseOpera ,itU see

,

changegiven
For the action choices (F, O), Given that Pat chooses Opera (O), if Chris unilaterally
changes his action to Football (F), he will receive 2 rather than 0. There is a profitable
unilateral deviation for Chirs. (F, O) is not a pure Nash equilibrium.
I

For the action2

receive rather
choices (F, F), Given O
, There isa
that Pat chooses F, if Chris unilaterally profitable
changes his
action to O, he will receive 0 rather than 2. There is a no profitable unilateral deviation
milateral deviation to Chris [F 0) is
for Chirs. Similarly, given that Chirs chooses F, if Pat unilaterally .changes her action to
,
O, she will receive 0 rather than 1. Thus, there is no profitable unilateral deviation for
not pre hash epilibrium
a (F, F) is a pure Nash equilibrium.
Pat.

his
IF F) .
Pat Choose F ,
if Chris changes
will receive O rather 2
action to 0 ,
he
chris
unaleral decisin for
profitable
.

There is no

Given Chris Choose F , Pat change it decision

she will receive o rather than 1 ,
to 0 ,

literal decisive
There is no profitable
E . F) is a nush equilibrium
"Pure
14
2.3.5.2 Business Application: Location Choice Game

The method of no profitable unilateral deviation is helpful in determining the Nash

equilibrium in games such as location choice. Let’s consider the location choice game
below.

There are 9 blocks in a street and each block has 10 customers residing in it.

Block 1 Block 2 Block 3 Block 4 Block 5 Block 6 Block 7 Block 8 Block 9

There are two street vendors, A and B, selling ice-cream (of identical taste and quality)
from a mobile cart. Price of a unit of ice-cream is fixed at $1. Each consumer buys ice-
cream from the nearest vendor. Each vendor chooses a location to maximize the
number of customers buying from him. Therefore, the action set for each vendor is:
{Block 1, Block 2, Block 3, Block 4, Block 5, Block 6, Block 7, Block 8, Block 9}.
Let 𝐿 and 𝐿 be the block chosen by vendor A and B respectively.

Consider the case where 𝑳𝑨 < 𝑳𝑩 .

A B
Block 1 Block 2 Block 3 Block 4 Block 5 Block 6 Block 7 Block 8 Block 9

 Suppose 𝐿 = 6 and 𝐿 = 9.
 Customers in Block 1 to 7 will buy ice-cream from A and customers in Block 8 and
9 will buy ice-cream from B.
 A has 70 customers while B has 20 customers only.
 B has an incentive to change its location unilaterally. If B moves to Block 8, he
gets 25 customers (assuming half of the customers in Block 7 will buy from B and
the other half will buy from A).
A B B
Block 1 Block 2 Block 3 Block 4 Block 5 Block 6 Block 7 Block 8 Block 9

 𝐿 = 6 and 𝐿 = 9 is not a Nash equilibrium.

B * LB
. K LA


~
You can try other combinations of 𝐿 and 𝐿 . As long as 𝑳𝑨 < 𝑳𝑩 , there is a
profitable unilateral deviation. it
Similarly, no equilibrium with 𝑳𝑨 > 𝑳𝑩 .
-

15
Now, consider 𝐿 = 𝐿 . However, there are 9 possibilities, e.g., 𝐿 = 𝐿 = 1, 𝐿 =
𝐿 = 2, etc.

Consider the case where 𝑳𝑨 = 𝑳𝑩 .

A, B
Block 1 Block 2 Block 3 Block 4 Block 5 Block 6 Block 7 Block 8 Block 9

 Suppose 𝐿 = 1 and 𝐿 = 1.
 Since they are located at the same block, each of them has 45 customers.
 However, both have the incentive to change location unilaterally.
 Suppose A moves to Block 2 unilaterally. After the move, it gets 80 customers.

A, B A
Block 1 Block 2 Block 3 Block 4 Block 5 Block 6 Block 7 Block 8 Block 9

 𝐿 = 𝐿 = 1 is not an equilibrium.
 The Nash equilibrium is 𝒙𝑨 = 𝒙𝑩 = 𝟓. Why?

A, B
Block 1 Block 2 Block 3 Block 4 Block 5 Block 6 Block 7 Block 8 Block 9

 Suppose A moves to Block 6 or Block 4 unilaterally. After the move, it gets 45

customers. There is no profitable unilateral deviation. The same happens for B.

16
2.4 Applications to Oligopoly

2.4.1 The Cournot Game

The Cournot model explains how firms behave if they simultaneously choose how
much they produce. In the basic Cournot model, we assume that:
F
1. O-
There are two firms (i.e., duopoly) and no others can enter the market.
2. The firms have identical costs.
3. The firms sell identical products.
4. The firms set their quantities simultaneously.

Let the total industry output be given by

𝑄 =𝑞 +𝑞
In other words, there are two firms in the industry, Firm 1 and Firm 2. Firm 1 chooses
𝒒𝟏 and Firm 2 chooses 𝒒𝟐 . The sum of their output 𝑸 = 𝒒𝟏 + 𝒒𝟐 is the total
-
industry output.

The market demand function is given by:

𝑃 = 50 − 2𝑄
Assume each firm has the same cost functions which are given by:
𝐶 = 10 + 2𝑞 , 𝑖 = 1,2
This means the cost function for Firm 1 is:
𝐶 = 10 + 2𝑞
>
-

And the cost function for Firm 2 is:

𝐶 = 10 + 2𝑞
-

Let’s define the details of the Cournot game.

 Players: Two firms.
 Strategies: Each firm chooses an output level [0, ∞). However, we must ensure
a positive price: 𝑃 = 𝑎 − 𝑏𝑄 > 0 ⟹ Q < a/b . Therefore, the strategy set of
each firm should be [0, 𝑎/𝑏).
 Payoffs: Each firm maximizes profits.
 Type of Game: Complete information & Static & Non-cooperative.
 Equilibrium Concept: We use the Nash equilibrium to solve the game.

17
2.4.1.1 Solving the Cournot Game

Each firm maximizes profits.

𝑃𝑟𝑜𝑓𝑖𝑡 = 𝑅𝑒𝑣𝑒𝑛𝑢𝑒 − 𝐶𝑜𝑠𝑡
𝑅𝑒𝑣𝑒𝑛𝑢𝑒 = 𝑃𝑟𝑖𝑐𝑒 × 𝑄𝑢𝑎𝑛𝑡𝑖𝑡𝑦
a = 41 + 42
Profit of Firm 1: -
𝜋 = (50 − 2𝑄) × 𝑞 − (10 + 2𝑞 )
𝜋 = (50 − 2(𝑞 + 𝑞 )) × 𝑞 − (10 + 2𝑞 )
𝜋 = 50𝑞 − 2𝑞 − 2𝑞 𝑞 − 10 − 2𝑞
𝜋 = 48𝑞 − 2𝑞 𝑞 − 2𝑞 − 10

Profit of Firm 2:
𝜋 = (50 − 2𝑄) × 𝑞 − (10 + 2𝑞I) -RP2 (firm2)
𝜋 = (50 − 2(𝑞 + 𝑞 )) × 𝑞 − (10 + 2𝑞 )
𝜋 = 50𝑞 − 2𝑞 𝑞 − 2𝑞 − 10 − 2𝑞 - Cost1 Same
· i

𝜋 = 48𝑞 − 2𝑞 𝑞 − 2𝑞 − 10

To solve the profit maximization problem, we first calculate the partial derivative:

and then set = 0. This is called the first-order condition (FOC).

↓ 10

22
For Firm 1:
𝜕𝜋
= 48 − 2𝑞 − 2(2)𝑞 = 0
𝜕𝑞
H
𝜕𝜋
𝜕𝑞
= 48 − 2𝑞 − 4𝑞 = 0
2(2)4
Then, we solve for 𝒒𝟏 .

𝑞 =
48 − 2𝑞
4
= 𝐵𝑅 (𝑞 ) F
We name it as 𝑩𝑹 - 𝟏 (𝒒𝟐 ): the best response function of Firm 1. 𝐵𝑅 (𝑞 ) means the
best response of Firm 1 is a function of 𝑞 . In other words, the best response of Firm
1 depends on 𝒒𝟐 .
-

-
depend 42

18
For Firm 2:
𝜕𝜋
= 50 − 2𝑞 − 2(2)𝑞 − 2 = 0
𝜕𝑞
Then, we solve for 𝑞 .
48 − 2𝑞
𝑞 = = 𝐵𝑅 (𝑞 )
4

Now, we have a system of equations with two equations two unknowns:

E
48 − 2𝑞
𝑞 = … (1)
4
48 − 2𝑞
𝑞 = … (2)
4
Substitute 𝑞 from the second equation to the first to get:
48 − 2𝑞
48 − 2( )
𝑞 = 4
4
As you can see, we now have one equation with one unknown only. We can solve for
𝑞 :
𝑞∗ = 8
Then, substitute 𝑞 ∗ into (2) and solve for 𝑞 . We get:
𝑞∗ = 8 # 4142
Therefore, the Nash equilibrium is 𝒒𝟏∗ = 𝒒∗𝟐 = 𝟖.

Profit for firm 1:

subin
𝜋 = 48𝑞 − 2𝑞 𝑞 − 2𝑞 − 10
𝜋 = 48(8) − 2(8)(8) − 2(8 ) − 10
profite
𝜋 = 118

Profit for firm 2 is the same as firm 1:

𝜋 = 118

19
 2.4.1.2 ICE
Consider the Cournot Game below. There are two firms in the industry, Firm 1 and
Firm 2. Firm 1 chooses 𝑞 and Firm 2 chooses 𝑞 . The sum of their output 𝑄 = 𝑞 +
𝑞 is the total industry output. The market demand function is given by:
𝑃 = 130 − 𝑄
Assume each firm has the same cost functions which are given by:

PX4 EE) cost 𝐶 = 10𝑞 , 𝑖 = 1,2

(a) Write down the profit function for each firm (𝜋 ).

(b) Derive the best response function for each firm (𝐵𝑅 (𝑞 ) and 𝐵𝑅 (𝑞 )).
(c) Find the Nash equilibrium.
(d) What is the equilibrium profit for each firm?

as
(a) :
Firm Firm
1
1:
Tr ,
(130 -1) 41 𝜋 -

1841
= (130 − 𝑄)𝑞 − 10𝑞

F1 = 12041
-

44 4142
𝜋 = (130 − 𝑞 − 𝑞 )𝑞 − 10𝑞
-

𝜋 = 120𝑞 − 𝑞 − 𝑞 𝑞
Firmh
120 p2-42𝜋 4142
Firm 2:
=
-

irz = 120𝑞 − 𝑞 − 𝑞 𝑞
(b)
b) FOC of Firm 1:

FirmI120-24
𝜕𝜋
= 120 − 2𝑞 − 𝑞 = 0
-𝜕𝑞↑
120 − 𝑞

/7542
𝑞 = = 𝐵𝑅 (𝑞 )
41 =
= BR2 , 24
Similarly
F

EEG-P-BR
120 − 𝑞
𝑞 = = 𝐵𝑅 (𝑞 )
2
(c)
120 − 𝑞
120 −
𝑞 = 2 → 𝑞 = 40

2412K
2
𝑞 = 40 = 40
(d) 41
Profit:
Wit
𝜋 = 𝜋 = 1600

↓) 20

profit i = Tvz = 1600

2.4.2 The Bertrand Game in Differentiated Products

Assume that there are two firms, 1 and 2, producing different brands of soft drink.
Since each firm produce a differentiated product, they face a different demand curve

T
and have market power to set the price.

Firm 1 has a demand function, 𝑞 = 100 − 2𝑝 + 𝑝 and cost function 𝐶 = 10𝑞

(no fixed cost). Firm 2 has a demand function, 𝑞 = 120 − 4𝑝 + 2𝑝 and a cost
function 𝐶 = 10𝑞 (no fixed cost).

What is the Nash equilibrium when firms choose their prices simultaneously?

Each firm maximizes profits.

𝑃𝑟𝑜𝑓𝑖𝑡 = 𝑅𝑒𝑣𝑒𝑛𝑢𝑒 − 𝐶𝑜𝑠𝑡
𝑅𝑒𝑣𝑒𝑛𝑢𝑒 = 𝑃𝑟𝑖𝑐𝑒 × 𝑄𝑢𝑎𝑛𝑡𝑖𝑡𝑦

The profit function of Firm 1:

𝜋 =𝑝 𝑞 −𝐶
𝜋 = 𝑝 𝑞 − 10𝑞
𝜋 = 𝑝 (100 − 2𝑝 + 𝑝 ) − 10(100 − 2𝑝 + 𝑝 )
𝜋 = 120𝑝 − 2𝑝 + 𝑝 𝑝 − 10𝑝 − 1000
FOC:
𝜕𝜋
= 120 − 4𝑝 + 𝑝 = 0
𝜕𝑝
Solve for 𝑝 :
𝑝
𝑝 = 30 + … (1)
4
This is the best response function of Firm 1.

The profit function of Firm 2:

𝜋 =𝑝 𝑞 −𝐶
𝜋 = 𝑝 𝑞 − 10𝑞
𝜋 = 𝑝 (120 − 4𝑝 + 2𝑝 ) − 10(120 − 4𝑝 + 2𝑝 )
𝜋 = 160𝑝 − 4𝑝 + 2𝑝 𝑝 − 20𝑝 − 1200
FOC:
𝜕𝜋
= 160 − 8𝑝 + 2𝑝 = 0
𝜕𝑝

21
Solve for 𝑝 :
𝑝
𝑝 = 20 + … (2)
4
This is the best response function of Firm 2.

Substitute 𝑝 from the second equation to the first to get:

𝑝
20 + 4
𝑝 = 30 +
4
∗
𝑝 = 37.33
∗
Then, substitute 𝑝 into (2) and solve for 𝑝. We get:
𝑝∗ = 29.33

Therefore, the Nash equilibrium is 𝑝∗ = 37.33, 𝑝∗ = 29.33.

2.4.3 Collusion Outcome in The Cournot Game

Wit like a single fire
~
Up to now we have considered two non-cooperative games: Cournot and Bertrand.
Now, we consider a cooperative outcome in the Cournot Game.

There are two firms in the industry, Firm 1 and Firm 2. Firm 1 chooses 𝑞 and Firm 2
chooses 𝑞 . The sum of their output 𝑄 = 𝑞 + 𝑞 is the total industry output. The
market demand function is given by:
𝑃 = 50 − 2𝑄
Assume each firm has the same cost functions which are given by:
𝐶 = 10 + 2𝑞 , 𝑖 = 1,2

Cooperation means the firms maximize the joint profit 𝜋 + 𝜋 by choosing the
total industry output level 𝑄. In other words, they act as a single firm.

Profit of Firm 1:
𝜋 = 𝑃𝑞 − 𝐶

Profit of Firm 2:
TV1 + ir L
𝜋 = 𝑃𝑞 − 𝐶 MAX

22
 * Tri
iv)
.

*
Joint profit:
𝜋 + 𝜋 = 𝑃𝑞 − 𝐶 + 𝑃𝑞 − 𝐶
𝜋 + 𝜋 = 𝑃𝑞 − 10 − 2𝑞 + 𝑃𝑞 − 10 − 2𝑞
𝜋 + 𝜋 = 𝑃𝑞 + 𝑃𝑞 − 2𝑞 − 2𝑞 − 20
↳= 41 42
+
𝜋 + 𝜋 = 𝑃(𝑞 + 𝑞 ) − 2(𝑞 + 𝑞 ) − 20 .

Using 𝑄 = 𝑞 + 𝑞

p = 50 - 20>. - 𝜋 + 𝜋 = 𝑃𝑄 − 2𝑄 − 20
𝜋 + 𝜋 = (50 − 2𝑄)𝑄 − 2𝑄 − 20
demand
sub market
𝜋 + 𝜋 = 50𝑄 − 2𝑄 − 2𝑄 − 20
𝜋 = 48𝑄 − 2𝑄 − 20
into (w)-it
joint / ,

41 +
FOC: 42
𝜕𝜋
= 48 − 4𝑄 = 0
𝜕𝑄
𝑄 ∗ = 12

Since both firms are identical, a natural choice for the outcome is equal shares: 𝑞 =
6, 𝑞 = 6.
-
12
m
The joint profit is: (
𝜋 = 48𝑄 − 2𝑄 − 20
𝜋 O
= 48(12) − 2(12 ) − 20 = 268

8
Assume the firms share the joint profit equally. Each firm will get 134.
-

Advantage of coropate .

23
2.4.4 Back to The Prisoners’ Dilemma

Now, let’s gather the payoffs from the non-cooperative and the cooperative Cournot.
The following is the normal-form representation.
Firm 2
Not cooperate (NC) Cooperate (C)
𝑞 =8 𝑞 =6
Firm 1 Not cooperate (NC) (118, 118) (??, ??)
𝑞 =8
Cooperate (C) (??, ??) (134, 134)
𝑞 =6

We still lack the payoffs of (C, NC) and (NC, C). Luckily, we know the profit function of
both firms:
𝜋 = 48𝑞 − 2𝑞 𝑞 − 2𝑞 − 10
𝜋 = 48𝑞 − 2𝑞 𝑞 − 2𝑞 − 10

Sub. 𝑞 = 6 and 𝑞 = 8 into the profit functions:

𝜋 = 48(6) − 2(6)(8) − 2(6 ) − 10 = 110
𝜋 = 48(8) − 2(6)(8) − 2(8 ) − 10 = 150

Sub. 𝑞 = 8 and 𝑞 = 6 into the profit functions:

𝜋 = 150
𝜋 = 110

Then, we can update the table.

Firm 2
Not cooperate (NC) Cooperate (C)
𝑞 =8 𝑞 =6
Firm 1 Not cooperate (NC) (118, 118) (150, 110)
𝑞 =8
Cooperate (C) (110, 150) (134, 134)
𝑞 =6

Hello! Is it you again? Yes, the Prisoners’ Dilemma! The pure Nash equilibrium is (NC,
NC).

24