BE

510 Business Economics 1 - Autumn 2021

Problem Set 8

1. Consider the following (incomplete) game trees.

B 𝑞!
1−
𝐼! A1
B 𝑝 𝑞! B

A N

B 1
− 𝑞 "
𝑝 1− B
𝐼" A2
𝑞"
B

(a) In the first game player A moves first. As indicated in the graph, her strategy is to let a die
throw decide on the action she takes. For example, if the outcome is 2, 3 or 4, she takes
the second action from top. She will never take the fourth action (the bottom one).
Suppose now that player B finds himself in information set 𝐼! . Given that A has used the
die throw strategy, what probability should B attach to being located in the first decision
node? What is the probability of the third decision node if he finds himself in 𝐼" ?
(b) In the second game Nature moves first and randomly decides on a type for player A (type
1 with probability 𝑝, type 2 with probability 1 − 𝑝). Subsequently, A uses a mixed strategy
that depends on her type (𝑞! versus 𝑞" ). Given this information, what probabilities should
B attach to each of the two decision nodes in his information set if he finds himself there?

2. Two firms share a common resource, and their payoffs depend on the total usage of the
resource. Firm 𝑖’s usage is denoted by 𝑥# . Firm 1 seeks to maximize
𝜋! = 𝑥! (36 − 𝑥! − 𝑥" ).
Firm 2, on the other hand, seeks to maximize either
𝜋"$ = 𝑥"$ (36 − 𝑥! − 𝑥"$ ) or 𝜋"% = 𝑥"% (24 − 𝑥! − 𝑥"% ).
Firm 2 knows its type but firm 1 doesn’t. It is common knowledge that firm 1 behaves as
described and that each type of firm 2 occurs with a 50-50 chance.
(a) Find the pure-strategy Bayes-Nash equilibrium of this simultaneous-move game.
(b) Would the firms prefer it if firm 2’s payoff function was common knowledge?
 BE 510 Business Economics 1 - Autumn 2021

3. Reconsider Problem Set 4 – Question 2 which revolved around the following setting.

A wants to meet B: Bob A wants to avoid B: Bob

Ballet Football Ballet Football
Ballet 3 , 2 1 , 1 Ballet 1 , 2 3 , 1
Alice Alice
Football 0 , 0 2 , 3 Football 2 , 0 0 , 3

Recall that Bob is uncertain whether Alice wants to meet him. Suppose it is common
knowledge that the probability of Alice wanting to meet him is 0.5. Find all mixed-strategy
Bayes-Nash equilibria.

 BE 510 Business Economics 1 - Autumn 2021

Problem Set 8 - Solutions

No guarantees for correctness. If you find errors in the proposed solutions, please let us know.

1. Consider the following (incomplete) game trees.

B 𝑞!
1−
𝐼! A1
B 𝑝 𝑞! B

A N

B 1
− 𝑞 "
𝑝 1− B
𝐼" A2
𝑞"
B

(a) In the first game player A moves first. As indicated in the graph, her strategy is to let a die
throw decide on the action she takes. For example, if the outcome is 2, 3 or 4, she takes
the second action from top. She will never take the fourth action (the bottom one).
Suppose now that player B finds himself in information set 𝐼! . Given that A has used the
die throw strategy, what probability should B attach to being located in the first decision
node? What is the probability of the third decision node if he finds himself in 𝐼" ?
Being located in information set 𝐼! player B knows that player A must have taken either
the first or the second action. Given A’s strategy the probability for the top decision node
at this stage is given by the ratio of the probability of the first action to the total probability
of 𝐼! :
Pr{Action! } Pr{Action! } 1⁄6 1
Pr{Action! |𝐼! } = = = =
Pr{𝐼! } Pr{Action! } + Pr{Action" } 1⁄6 + 3⁄6 4
Being located in information set 𝐼" player B knows of course that player A took the third
action (given the die throw strategy). Formally:
Pr{Action& } Pr{Action& } 2⁄6
Pr{Action& |𝐼" } = = = = 1
Pr{𝐼" } Pr{Action& } + Pr{Action' } 2⁄6 + 0
(b) In the second game Nature moves first and randomly decides on a type for player A (type
1 with probability 𝑝, type 2 with probability 1 − 𝑝). Subsequently, A uses a mixed strategy

1
Problem Set 8 - Solutions BE 510 Business Economics 1 - Autumn 2021

that depends on her type (𝑞! versus 𝑞" ). Given this information, what probabilities should
B attach to each of the two decision nodes in his information set if he finds himself there?
Using the same principle as before we get
𝑝𝑞!
Pr{First node} =
𝑝𝑞! + (1 − 𝑝)(1 − 𝑞" )
and
(1 − 𝑝)(1 − 𝑞" )
Pr{Second node} =
𝑝𝑞! + (1 − 𝑝)(1 − 𝑞" )

2. Two firms share a common resource… Firm 1 seeks to maximize 𝜋! = 𝑥! (36 − 𝑥! − 𝑥" ). Firm
2, on the other hand, seeks to maximize either
𝜋"$ = 𝑥"$ (36 − 𝑥! − 𝑥"$ ) or 𝜋"% = 𝑥"% (24 − 𝑥! − 𝑥"% ).
Firm 2 knows its type but firm 1 doesn’t. It is common knowledge that firm 1 behaves as
described and that each type of firm 2 occurs with a 50-50 chance.
(a) Find the pure-strategy Bayes-Nash equilibrium of this simultaneous-move game.
Firm 1 knows that it is dealing either with firm 2A or with firm 2B. Each of these will use
its own strategy, that is firm 2A will choose some 𝑥"$ and firm 2B will choose a different
𝑥"% . Hence, firm 1’s expected payoff is
1 1 𝑥"$ + 𝑥"%
𝜋! = 𝑥! (36 − 𝑥! − 𝑥"$ ) + 𝑥! (36 − 𝑥! − 𝑥"% ) = 𝑥! E36 − 𝑥! − F
2 2 2
Based on this, firm 1’s best response function is
𝑥"$ + 𝑥"%
𝑥! = 18 −
4
Firm 2A’s and firm 2B’s payoffs are as described in the question and so their best response
functions are
1 1
𝑥"$ = 18 − 𝑥! and 𝑥"% = 12 − 𝑥!
2 2
Hence, we obtain an equation system with 3 equations and 3 unknowns. Solving this
yields 𝑥!∗ = 14, 𝑥"$
∗ ∗
= 11 and 𝑥"% = 5 as our equilibrium solution.
(b) Would the firms prefer it if firm 2’s payoff function was common knowledge?
Now consider the two Nash equilibria that emerge from firm 1’s interaction with firm 2A
and with firm 2B in isolation. Firm 1’s first-order condition could be written as
𝑥! = 36 − 𝑋
where 𝑋 = 𝑥! + 𝑥" . Firm 2’s first-order condition could be written as

2

Problem Set 8 - Solutions BE 510 Business Economics 1 - Autumn 2021

𝑥" = 𝑎 − 𝑋
where 𝑎 = 36 for firm 2A and 𝑎 = 24 for firm 2B. Summing up the two left-hand-sides
and the two right-hand-sides we get
36 + 𝑎
𝑋 = 36 + 𝑎 − 2𝑋 and hence 𝑋 ∗ =
3
which implies
36 + 𝑎 𝑎 36 + 𝑎 2
𝑥!∗ = 36 − = 24 − and 𝑥"∗ = 𝑎 − = 𝑎 − 12
3 3 3 3
The payoffs would be
𝑎 𝑎 2 𝑎 "
𝜋! = M24 − N O36 − M24 − N − E 𝑎 − 12FP = M24 − N
3 3 3 3

and
"
2 𝑎 2 2
𝜋" = E 𝑎 − 12F O𝑎 − M24 − N − E 𝑎 − 12FP = E 𝑎 − 12F
3 3 3 3

Therefore, under complete information the payoffs are:

Profit for firm 1 Profit for firm 2

Firm 2 is of type A: 36 " 2 "

(𝑎 = 36) E24 − F = 144 E × 36 − 12F = 144

3 3

Firm 2 is of type B: 24 " 2 "

(𝑎 = 24) E24 − F = 256 E × 24 − 12F = 16

3 3
𝟏𝟒𝟒 + 𝟐𝟓𝟔 𝟏𝟒𝟒 + 𝟏𝟔
In expectation: = 𝟐𝟎𝟎 = 𝟖𝟎
𝟐 𝟐

Under incomplete information the payoffs are (from above):

Profit for firm 1 Profit for firm 2

Firm 2 is of type A: 𝑥!∗ (36 − 𝑥!∗ − 𝑥"$

∗ ) ∗ (36
= 154 𝑥"$ − 𝑥!∗ − 𝑥"$
∗ )
= 121

Firm 2 is of type B: 𝑥!∗ (36 − 𝑥!∗ − 𝑥"%

∗ )
= 238 ∗ (24
𝑥"% − 𝑥!∗ − 𝑥"%
∗ )
= 25

𝟏𝟓𝟒 + 𝟐𝟑𝟖 𝟏𝟐𝟏 + 𝟐𝟓

In expectation: = 𝟏𝟗𝟔 = 𝟕𝟑
𝟐 𝟐

Thus, in expectation both firms would prefer complete information for firm 1. When we
look at this in more detail, however, the answer is less clear: Firm 1 would prefer to have

3

Problem Set 8 - Solutions BE 510 Business Economics 1 - Autumn 2021

the information if firm 2 was of type B but would prefer to remain ignorant (and be known
to be ignorant!) if firm 2 was of type A. For firm 2 it is the other way around.

3. Reconsider Problem Set 5 – Question 2 which revolved around the following setting.

A wants to meet B: Bob A wants to avoid B: Bob

Ballet Football Ballet Football
Ballet 3 , 2 1 , 1 Ballet 1 , 2 3 , 1
Alice Alice
Football 0 , 0 2 , 3 Football 2 , 0 0 , 3

Recall that Bob is uncertain whether Alice wants to meet him. Suppose it is common
knowledge that the probability of Alice wanting to meet him is 0.5. Find all mixed-strategy
Bayes-Nash equilibria.
Alice’s pure strategies are BB, BF, FB, and FF (first letter for the action taken by the Alice who
wants to meet Bob; second letter for the action taken by the Alice who wants to avoid him).
Bob’s pure strategies are Ballet and Football.
Consider the possibility that only one player randomizes. However, looking at the payoffs it is
clear that whenever Bob takes one action for sure, one of Alice’s four pure strategies is the
unique best response: The Alice who wants to meet Bob best responds by choosing the same
location as he does, and the Alice who wants to avoid him best responds by choosing the other
location. Conversely, whenever Alice picks one of her four strategies, either Ballet or Football
is the unique best response for Bob: If she picks BB or FF, Bob best responds by choosing the
same location. If she chooses BF or FB, he maximizes his payoff by taking Football (given the
50-50 chance for each type of Alice and given that he prefers Football).
Thus, given that one player doesn’t randomize the other shouldn’t randomize either. The only
remaining possibility is that both players randomize simultaneously. Thus, suppose that Bob
chooses Football with probability 𝑝. What are the implications for Alice? A mixed strategy
equilibrium requires that Alice becomes indifferent between some of her options.
Consider Alice 1 who wants to meet Bob. Alice 1 is indifferent between Ballet and Football
when Bob chooses Ballet with probability 1⁄4: (1⁄4) × 3 + (3⁄4) × 1 = (3⁄4) × 2. Alice 2 is
indifferent between Ballet and Football when Bob chooses Ballet with probability 3⁄4 :
(3⁄4) × 1 + (1⁄4) × 3 = (3⁄4) × 2. See also Problem Set 5, Question 2. Each of these cases
may possibly lead to an equilibrium but note that the two types of Alice cannot both be
indifferent at the same time.
Case 1: If Bob chose Ballet with probability 1⁄4, Alice 1 would become indifferent between
Ballet and Football, but Alice 2 would strictly prefer Ballet. Suppose that Alice 1 chooses Ballet
with probability 𝑞 and Football with probability 1 − 𝑞. Alice 2, in contrast, must choose Ballet
with probability 1.

4

Problem Set 8 - Solutions BE 510 Business Economics 1 - Autumn 2021

What is Bob’s overall chance of meeting Alice at the stadium? Answer: It’s the probability that
she is of type 1 times the probability that Alice 1 picks Football: (1⁄2)(1 − 𝑞).
What is Bob’s overall chance of meeting Alice at the ballet performance? Answer: He meets
her either when she is of type 2 (50% chance), or when she is of type 1 and picks Ballet (50%
chance times q): (1⁄2) + (1⁄2)𝑞 = (1 + 𝑞)⁄2.
Consequently, in this scenario Bob’s expected utility from choosing Ballet is
1+𝑞
𝜋Bob (B) = 2 × = 1 + 𝑞
2
and his expected utility from choosing Football is
1−𝑞 1−𝑞
𝜋Bob (F) = 3 × + 1 × E1 − F = 2 − 𝑞.
2 2
To have a mixed-strategy equilibrium Bob must be indifferent between Ballet and Football.
This occurs when 𝑞 = 1⁄2. Thus, there is a mixed-strategy Bayes-Nash Equilibrium in
which Bob chooses Ballet with probability 𝟏⁄𝟒 and Football with probability 𝟑⁄𝟒, and
Alice chooses BB, BF, FB, and FF with probabilities 𝟏⁄𝟐, 𝟎, 𝟏⁄𝟐 and 0, respectively.

Case 2: If Bob chose Ballet with probability 3⁄4, Alice 2 would become indifferent between
Ballet and Football, but Alice 1 would strictly prefer Ballet. Suppose that Alice 2 picks Ballet
with probability 𝑞. Then Bob’s chance of meeting Alice at the stadium is the probability that
she is of type 2 times the probability that Alice 2 picks Football: (1⁄2)(1 − 𝑞), same as before.
Bob meets Alice at the ballet performance either when she is of type 1 (50% chance), or when
she is of type 2 and picks Ballet: (1⁄2) + (1⁄2)𝑞 = (1 + 𝑞)⁄2. This is also as before. Bob is
again indifferent between Ballet and Football when 𝑞 = 1⁄2 . Thus, there is a mixed-
strategy Bayes-Nash Equilibrium in which Bob chooses Ballet with probability 𝟑⁄𝟒 and
Football with probability 𝟏⁄𝟒, and Alice chooses BB, BF, FB, and FF with probabilities
𝟏⁄𝟐, 𝟏⁄𝟐, 𝟎 and 0, respectively.