Bayesian Nash Equilibrium

We have already seen that a strategy for a player in a

game of incomplete information is a function that speci-
fies what action or actions to take in the game, for every
possible type of that player.

A Bayesian Nash Equilibrium is a Nash equilibrium of

this game (in which the strategy set is the set of action
functions).

There are two ways of finding a pure-strategy Bayesian

Nash Equilibrium (BNE).

Method 1. This method works directly on the Bayesian

normal form representation, which is most easily done by
converting the game into the corresponding payoff ma-
trix. Simply find the Nash equilibria from the payoff ma-
trix.
This method computes expected payoffs from an ex ante
perspective, before the players learn their types. Notice
that, if the set of actions available or the set of possible
types is infinite, we cannot construct the payoff matrix
so method 1 will not work.

Here is the payoff matrix for the Entry Game with Cost
Uncertainty, with best responses marked with a star.

firm 2
E N
E LE H − 14 , − 14 3 ∗, 0∗
∗ 4
firm 1 E LN H 0∗, 14 1, 0
2
∗
N LE H − 14 , 14 1, 0
4
∗
N LN H 0∗, 34 0, 0

There are three BNE: (E LE H , N), (N LN H , E), and

(E LN H , E).
Here is the payoff matrix for the Gift Game, with best
responses marked with a star for the case in which player
1 is more likely to be a friend p > 12 .

player 2
A R
NF NE 0, 0∗ 0∗, 0∗
player 1 N F GE 1 − p, p − 1 p − 1, 0∗
GF N E p, p∗ −p, 0
GF GE 1∗, 2p − 1∗ −1, 0

Thus, when p > 12 , the BNE are (N F N E , R) and (GF GE , A).

Here is the payoff matrix for the Gift Game, with best
responses marked with a double star for the case in which
player 1 is more likely to be an enemy p < 12 .

player 2
A R
NF NE 0, 0∗∗ 0∗∗, 0∗∗
player 1 N F GE 1 − p, p − 1 p − 1, 0∗∗
GF N E p, p∗∗ −p, 0
GF GE 1∗∗, 2p − 1 −1, 0∗∗

Now the game has only one BNE, (N F N E , R). No

matter what p is, it is an equilibrium for player 1 never
to offer because the gift will be refused. There is also a
BNE in which player 2 accepts and both types of player
1 offer, but only if type F is more likely than type E.
Method 2. This method for finding the BNE converts
the game into an equivalent "bigger" game in which the
different types of each player are treated as separate play-
ers. The payoff to player (i, ti) is the expected payoff of
player i, conditional on being type ti.

Any NE of the bigger game is a BNE of the original game,

and vice versa:

The payoff in the Bayesian normal-form matrix is the

summation over all types of the probability of a type mul-
tiplied by the expected payoff conditional on that type.

=⇒If player i is best responding in the original game, then

it is impossible to increase his payoff conditional on any
of his types (or else this summation would be higher), so
each player (i, ti) must be best responding in the bigger
game.
⇐=If each player (i, ti) is best responding in the big-
ger game, then it is impossible to increase his ex ante
expected payoff in the original game by choosing a differ-
ent strategy function ai(ti). (The only way to achieve a
higher summation of probability of ti multiplied by util-
ity conditional on ti is to increase at least one term in
the sum, which is impossible if each player (i, ti) is best
responding in the bigger game.)

Method 2 is often easier than Method 1, especially when

players have an infinite number of possible actions (Cournot
game with cost uncertainty) or an infinite number of types
(auction with a continuous distribution of possible valu-
ations for the object being auctioned).

Let us apply Method 2 to the Entry Game with Cost

Uncertainty, and see that we get the same answer as in
Method 1.
 firm 2
E N
firm 1 E −c1, −c2 1 − c1, 0
N 0, 1 − c2 0, 0

In the bigger game, there are three players, (1, H), (1, L),
and 2. Let us find the NE of this game.

Can there be a NE in which firm 2 chooses E? Player

(1, H) receives a payoff of 0 by choosing N H , and a
payoff of − 12 from choosing E H , so the best response
is N H . Player (1, L) receives a payoff of 0 by choosing
N L, and a payoff of 0 from choosing E L, so both N L
and E L are best responses.

Firm 2’s choice of E is the best response to the pro-

file (N L, N H ) [because 34 > 0] and it is a best re-
sponse to the profile (E L, N H ) [because 14 > 0], so
both (N L, N H , E), and (E L, N H , E) are NE of the 3-
player game. Therefore, (N LN H , E), and (E LN H , E)
are BNE of the Bayesian normal-form game.
Can there be a NE in which firm 2 chooses N? Player
(1, H) receives a payoff of 0 by choosing N H , and a
payoff of 12 from choosing E H , so the best response is
E H . Player (1, L) receives a payoff of 0 by choosing
N L, and a payoff of 1 from choosing E L, so the best
response is E L.

Firm 2’s best response to the profile (E L, E H ) is N

[because 0 > − 14 ]. Therefore, (E L, E H , N) is a NE
of the 3-player game, and (E LE H , N) is a BNE of the
Bayesian normal-form game.
 Cournot Competition with Cost Uncertainty

Consider a (simultaneous move) Cournot game with the

inverse demand function

p = 1 − q1 − q2.

Firm 1’s production cost is zero.

Firm 2 has two possible types, L and H, each of which

occur with probability 12 .

A type L firm 2 has low marginal cost, 0, and a type H

firm 2 has high marginal cost, 14 .

In the Bayesian normal-form game, a strategy for firm 1

is a quantity, q1, and a strategy for firm 2 is a function
that specifies a quantity for each type, (q2L, q2H ).
Since every nonnegative quantity is a possible strategy for
firm 1 and every pair of nonnegative quantities is a possi-
ble strategy for firm 2, the resulting payoff matrix would
have an infinite number of rows and columns! Clearly
Method 1 will not be easy.

Under Method 2, we consider the three player game with

firm 1, firm 2L, and firm 2H. To find the NE, we com-
pute the best response functions for all three players and
solve the three equations for the three NE quantities.

Starting with firm 2L, its payoff function is

uL
2 = (1 − q1 − q L)q L.
2 2
Differentiating with respect to q2L, setting the expression
equal to zero, and solving for q2L, we can solve for firm
2L’s best response function.

∂uL2 = 0 = 1 − q − 2q L
1 2
∂q2L
L 1 − q1
BR2 (q1) = .
2
The payoff function of firm 2H is given by

H H H q2H
u2 = (1 − q1 − q2 )q2 − .
4

Differentiating with respect to q2H , setting the expression

equal to zero, and solving for q2H , we can solve for firm
2H’s best response function.

∂uH2 = 0 = 1 − q − 2q H − 1
1 2
∂q2H 4
3 q1
BR2H (q1) = − .
8 2
The payoff function for firm 1 is based on the expectation
that half of the time it is competing with firm 2L and
half of the time it is competing with firm 2H.

1 1
u1 = [(1 − q1 − q2L)q1] + [(1 − q1 − q2H )q1]
2 2
q2L q2H
= (1 − q1 − − )q1
2 2

Differentiating with respect to q1, setting the expression

equal to zero, and solving for q1, we can solve for firm
1’s best response function.

∂u1 q2L q2H

= 0 = 1 − 2q1 − −
∂q1 2 2
L H 1 q2L q2H
BR1(q2 , q2 ) = − − .
2 4 4
The Nash equilibrium is the solution to the following three
equations
1 q1
BR2L(q1) = − = q2L
2 2
3 q1
BR2H (q1) = − = q2H
8 2
L H 1 q2L q2H
BR1(q2 , q2 ) = − − = q1
2 4 4

To solve, substitute q2L from the first equation and q2H

from the second equation into the third equation, and
solve for q1.

1 1 q 3 q
− ( − 1 ) − ( − 1 ) = q1
2 8 8 32 8
9 q
+ 1 = q1
32 4
4 9 3
q1 = ( ) = .
3 32 8
Substituting q1 = 38 into the remaining two equations, we
have the Nash equilibrium strategy profile for this three
player game, q1 = 38 , q2L = 16
5 , qH = 3 .
2 16

Thus, the BNE for the original game with two players is
the following:
3
q1 =
8
L H 5 3
(q2 , q2 ) = ( , ).
16 16

For this game, firm 1’s quantity is a best response to the

average quantity selected by firm 2.
 Auction Markets

We will solve the games corresponding to various auction

rules under the following environment:

There is one indivisible object being sold.

The players are the n bidders. Each player has a valuation

for the object, with player i’s valuation denoted by vi.

If player i wins the auction and makes a payment, p, her

overall payoff is vi − p; if she does not win the auction
but she makes a payment, p, her overall payoff is −p.

We assume that each vi is independently drawn from the

uniform distribution over the unit interval [0, 1]. In other
words, all realizations between 0 and 1 are equally likely,
and knowing vi provides no information about the other
players’ valuations.
1. Sealed-bid, first-price auction.

The players simultaneously submit bids, where the bid

of player i is denoted by bi. Each player observes her
valuation (her type) before deciding what to bid, so a
strategy is a bid function, bi(vi).

The player submitting the highest bid wins the auction

and makes a payment equal to her bid.

Players who do not win the auction do not make a pay-

ment; their payoff is zero. In case of a tie for the highest
bid, someone is randomly selected as the winner.

What would you bid if n = 10 and your valuation is 0.6?

Notice that it does not make sense to bid more than your
valuation, because your payoff cannot be positive. Either
you lose the auction and receive zero, or win the auction
and receive a negative payoff.

In fact, players should bid less than their valuation so that

if they win the auction, their payoff is positive.

Finding the BNE is not easy. We will guess that there

is a symmetric BNE in which all players bid a constant
fraction of their valuation:

bi(vi) = avi
for some number a that is the same for all players and is
between zero and one.

Then we will use the condition that every type of every

player is best-responding to the other players by bidding
this way. This will allow us to solve for the value of a
that make this symmetric profile of bidding functions a
BNE.
Consider player j with valuation vj , and suppose that all
of the other players are bidding according to bi(vi) = avi.
Then the highest possible bid by one of the other players
(with vi = 1) is a, so there is no reason to bid more than
a.

If player j makes a bid of b, she wins the auction if and

only if all of the other bids are below b. Based on their
bidding functions, this happens if we have for all i 6= j,

avi < b, or
b
vi < .
a
Because of the uniform distibution, the probability that a
particular one of the other players has a valuation below
b is b .
a a

Then the probability that all of the other players have

valuations below ab , so that player j wins the auction, is
given by
µ ¶n−1
b
pr(j wins when bidding b) = .
a

This allows us to express player j’s payoff as a function

of her valuation and her bid, given the bidding functions
of the other players.
µ ¶n−1
b
uj = [vj − b]
a
We can now find the optimal bid for each type of player
j, by taking the derivative of uj with respect to b, setting
the expression equal to zero, and solving for b.

a1−n[(n − 1)bn−2(vj − b) + bn−1(−1)] = 0

(n − 1)(vj − b) − b = 0
(n − 1)vj − nb = 0
n−1
b = vj
n

Thus, if other players are bidding a constant fraction of

their valuations (no matter what the constant, a), the
best response of bidder j is to bid a constant fraction,
n−1 , of her valuation.
n

Therefore, we have a BNE if each player i uses the bidding

function,
n−1
bi(vi) = vi.
n
Recapping, in the sealed-bid first-price auction, the BNE
is for all players to use the bidding function, bi(vi) =
n−1 v .
n i

Players shade their bid below their valuation, to balance

the profit when they win against the risk of not winning.
The more players in the auction, the less they can afford
to shade their bid.

The BNE outcome is efficient, because the player with

the highest valuation always wins the auction.

When n = 2, players bid half their valuation, so when

the highest valuation is v, the player with that valuation
receives a payoff of v2 and the seller receives revenue of
v.
2
2. Sealed-bid, second-price auction.

The players simultaneously submit bids, where the bid

of player i is denoted by bi. Each player observes her
valuation (her type) before deciding what to bid, so a
strategy is a bid function, bi(vi).

The player submitting the highest bid wins the auction

and makes a payment equal to the second highest bid.

Players who do not win the auction do not make a pay-

ment; their payoff is zero. In case of a tie for the highest
bid, someone is randomly selected as the winner.

For example, if player 3 submits the highest bid of 0.78

and player 6 submits the second-highest bid of 0.62, then
player 3 wins the auction (driving home with the object)
and pays 0.62, while the other players do not receive the
object or make any payments.

What would you bid if n = 10 and your valuation is 0.6?

One could try to solve for the BNE of the sealed-bid
second-price auction the same way that we solved the
first-price auction, but there is a much easier way.

Notice that it is a weakly dominant strategy to bid your

valuation:

If the highest of the other players’ bids (call it v0) is

greater than vi, then bidding vi is a best response. Player
i loses, but changing her bid in order to win would require
her to bid more than v 0, in which case she would pay v 0,
which is more than her valuation.

If v 0 < vi, then bidding vi is a best response. Player i

wins the auction and receives a positive payoff. Changing
her bid while still winning does not change her payment,
v 0, and changing her bid to something below v 0 reduces
her payoff to zero.
The symmetric BNE has all players choosing the bidding
strategy bi(vi) = vi.

The BNE outcome is efficient, because the player with

the highest valuation always wins the auction.

When n = 2, since players bid their valuation, when

the highest valuation is v, the player with that valuation
wins the auction and makes a payment equal to the other
player’s valuation. The payment is uniformly distributed
over the interval from 0 to v, so the expected payment
is v2 . Thus, the winner receives an expected payoff of v2
and the seller receives expected revenue of v2 .

The expected payoff to the players and the expected rev-

enue to the seller is the same as in the first-price auction.
This result that payoffs do not depend on the auction
format (called revenue equivalence) extends to n players
and to valuation distributions other than uniform.
These sealed-bid auctions have dynamic counterparts.

The (independent private values) second-price sealed bid

auction is essentially equivalent to the ascending-price or
English auction you see in movies or on eBay.

The (independent private values) first-price sealed bid

auction is essentially equivalent to the descending-price
or Dutch auction.

The game changes significantly if payoffs have a "com-

mon value" component. That is, if the object’s worth to
me increases when I learn that the object is worth a lot
to you.

For example, if we are auctioning two tickets to the Buck-

eye game on the 50 yard line in the first row of C-deck,
then the bidders know their valuations, which would not
change if they learned that the other bidders had high
valuations. If instead the seat locations are not specified
and bidders’ types in part reflect information about where
the seats are, then learning your type tells me something
about what the seats would be worth to me.