Varian9e LecturePPTs Ch30

Chapter 30
Game
Applications
Nash Equilibrium
 In any Nash equilibrium (NE) each

player chooses a “best” response to
the choices made by all of the other
players.
 A game may have more than one NE.
Nash Equilibrium
 How can we locate every one of a

game’s Nash equilibria?
 If there is more than one NE, can we
argue that one is more likely to occur
than another?
Best Responses
 Think of a 2×2 game; i.e., a game

with two players, A and B, each with
two actions.
 A can choose between actions aA1
and aA2.
Best Responses
 B can choose between actions aB1

and aB2.
 There are 4 possible action pairs;
(aA1, aB1), (aA1, aB2), (aA2, aB1), (aA2, aB2).
 Each action pair will usually cause
different payoffs for the players.
Best Responses
 Suppose that A’s and B’s payoffs when
the chosen actions are aA1 and aB1 are
UA(aA1, aB1) = 6 and UB(aA1, aB1) = 4.
 Similarly, suppose that
UA(aA1, aB2) = 3 and UB(aA1, aB2) = 5
Best Responses
 UA(aA1, aB1) = 6 and UB(aA1, aB1) = 4

Best Responses

 If B chooses action aB1 then A’s best
response is ??
Best Responses

response is action aA1 (because 6 > 4).
Best Responses

response is ??
Best Responses

response is action aA2 (because 5 > 3).
Best Responses
 If B chooses aB1 then A chooses aA1.
 If B chooses aB2 then A chooses aA2.
 A’s best-response “curve” is
therefore
aA2 +
A’ s
best aA1 +
respons
e
a B1 a B2 B’ s
action
Best Responses

Best Responses

 If A chooses action aA1 then B’s best
response is ??
Best Responses

response is action aB2 (because 5 > 4).
Best Responses

response is ??.
Best Responses

response is action aB2 (because 7 > 3).
Best Responses
 If A chooses aA1 then B chooses aB2.
 B’s best-response “curve” is
therefore
aA2
A’ s
actio aA1
n
a B1 a B2 B ’ s best
response
Best Responses
 B’s best-response “curve” is
therefore
a A Notice that aB2 is a
2
strictly dominant
A’ s
aA1 action for B.
actio
n
a B1 a B2 B ’ s best
response
Best Responses & Nash Equilibria
How can the players’ best-response curves be

used to locate the game’ s Nash equilibria?
A’ s A’ s
B A
response choice
aA2 aA2 +
aA1 aA1 +
a B1 a B2 a B1 a B2
B’ s B’ s
choice response

used to locate the game’ s Nash equilibria? Pu
A’ s A’ s one curve on top of
B A
response choice the other.
aA2 aA2 +
aA1 aA1 +
a B1 a B2 a B1 a B2
B’ s B’ s
choice response

A’ s A’ s one curve on top of
B A
response choice the other.
aA2 aA2 +
aA1 aA1 +
a B1 a B2 a B1 a B2
B’ s B’ s
choice response

A’ s one curve on top of
response the other.
aA2 + Is there a Nash equilibrium?
aA1 +
a B1 a B2
B’ s
response

response the other.
Yes, (aA2, aB2). Why?
aA1 +
a B1 a B2
B’ s
response

response the other.
Yes, (aA2, aB2). Why?
aA1 + aA2 is a best response to aB2.
aB2 is a best response to aA2.
a B1 a B2
B’ s
response
Player B Here is the strategic

aB1 aB2 form of the game.
aA1 6,4 3,5

Player A
aA2 4,3 5,7
aA2 is the only best response to aB2.

aB2 is the only best response to aA2.

Is there a 2nd Nash
aA1 6,4 3,5
eqm.?
Player A
aA2 4,3 5,7


Is there a 2nd Nash
aA1 6,4 3,5
eqm.? No, because
Player A aB2 is a strictly
aA2 4,3 5,7 dominant action
for Player B.

Player B
aB1 aB2
aA1 6,4 3,5

Player A
aA2 4,3 5,7
Now allow both players to randomize (i.e., mix)

over their actions.
Player B
aB1 aB2 A1 is the prob. A
chooses action aA1.
aA1 6,4 3,5 B1 is the prob. B
Player A chooses action aB1.
aA2 4,3 5,7
Now allow both players to randomize (i.e., mix)

over their actions.
Player B
chooses action aA1.
aA2 4,3 5,7 Given B1, what
value of A1 is best
for A?
Player B
chooses action aA1.
a 2
A
4,3 5,7 Given  B
1 , what
EVA(aA1) = 6B1 + 3(1 - B1) = 3 for
+ 3A?B
1.
Player B
chooses action aA1.
aA2 4,3 5,7 Given B1, what
EVA(aA1) = 6B1 + 3(1 - B1) = 3 for
+ 3A? B
1.
EVA(aA2) = 4B1 + 5(1 - B1) = 5 - B1.
A1 is the prob. A chooses action aA1.

B1 is the prob. B chooses action aB1.
Given B1, what value of A1 is best for A?
EVA(aA1) = 3 + 3B1.
EVA(aA2) = 5 - B1.
3 + 3B1>= 5 - B
1 as  B>
=
1 ??
< <

EVA(aA1) = 3 + 3B1.
EVA(aA2) = 5 - B1.
3 + 3B1>= 5 - B
1 as  B>
=
1 ½.
< <

EVA(aA1) = 3 + 3B1.
EVA(aA2) = 5 - B1.
3 + 3B1>= 5 - B
1 as  B>
=
1 ½.
< <
A ’ s best response is:
aA1 if B1 > ½
aA2 if B1 < ½
aA1 or aA2 if B1 = ½

EVA(aA1) = 3 + 3B1.
EVA(aA2) = 5 - B1.
3 + 3B1>= 5 - B
1 as  B>
=
1 ½.
< <
A ’ s best response is:
aA1 (i.e. A1 = 1) if B1 > ½
aA2 (i.e. A1 = 0) if B1 < ½
aA1 or aA2 (i.e. 0  A1  1) if B1 =
A ’ s best response is: aA1 (i.e. A1 = 1) if B1

>½
aA2 (i.e. A1 = 0) if B1 <
A’ ½
s best
response
A1 aA1 or aA2 (i.e. 0  A1 
1) if1
B 1 = ½
0
0 ½ 1  B1
A ’ s best response is: aA1 (i.e. A1 = 1) if

 B1 > ½
aA2 (i.e. A1 = 0) if B1 < ½
A ’ s best aA1 or aA2 (i.e. 0  A1  1)
response

if 1
A
1 B 1 = ½
0
0 ½ 1  B1

>½
aA2 (i.e. A1 = 0) if B1 <
A’ ½s best
response
A1 aA1 or aA2 (i.e. 0  A1  1)
if 1
B 1 = ½
0
0 ½ 1  B1

>½
aA2 (i.e. A1 = 0) if B1 <
A’ ½
s best
response
A1 aA1 or aA2 (i.e. 0  A1 
1) if
1
This is A’ s best response
curve when players are
 B
1 =½ allowed to mix over their
actions.
0
0 ½ 1  B1
Player B
chooses action aA1.
aA2 4,3 5,7 Given A1, what
value of B1 is best
for B?
Player B
chooses action aA1.
a 2
A
4,3 5,7 Given  A
1, what
EVB(aB1) = 4A1 + 3(1 - A1) = 3 for
+ B?
A
1.
Player B
chooses action aA1.
a 2
A
4,3 5,7 Given  A
1, what
EVB(aB1) = 4A1 + 3(1 - A1) = 3 for
+ B?A
1.
EVB(aB2) = 5A1 + 7(1 - A1) = 7 - 2A1.

Given A1, what value of B1 is best for B?
EVB(aB1) = 3 + A1.
EVB(aB2) = 7 - 2A1.
3 + A1>= 7 - 2 A
1 as A>
=
1 ??
< <

Given A1, what value of B1 is best for B?
EVB(aB1) = 3 + A1.
EVB(aB2) = 7 - 2A1.
3 + A1 < 7 - 2A1 for all 0  A1  1.

EVB(aB1) = 3 + A1.
EVB(aB2) = 7 - 2A1.
3 + A1 < 7 - 2A1 for all 0  A1  1.
B ’ s best response is:
aB2 always (i.e. B1 = 0 always).
B ’ s best response is aB2 always (i.e. B1 = 0

always).
A1
1
This is B’ s best response
curve when players are
allowed to mix over their
actions.
0
0 ½ 1  B1
B ’ s best response
B A
A ’ s best response
A1 A1
1 1
0 B
0
0 ½ 1  1 0 ½ 1  B1
Is there a Nash equilibrium?
B A
A1 A1
1 1
0 B
0
0 ½ 1  1 0 ½ 1  B1
B A
A1 A1
1 1
0 B
0
0 ½ 1  1 0 ½ 1  B1
A ’ s best
responseA1
1
0
0 ½ 1  B1
B ’ s best
Is there a Nash equilibrium? Yes. Just one.

(A1, B1) = (0,0);
i.e. A chooses aA2 only
A ’ s best response & B chooses aB2 only.
A1
1
0
0 ½ 1  B1
B ’ s best
Player B Let’ s change the game

aB1 aB2
aA1 6,4 3,5

Player A
aA2 4,3 5,7
Player B Here is a new

aB1 aB2 2×2 game.
aA1 6,4 3,1

3,5
Player A
aA2 4,3 5,7
Player B Here is a new

aB1 aB2 2×2 game. Again
let A1 be the prob.
that A chooses aA1
aA1 6,4 3,1
and let B1 be the
Player A prob. that B chooses
aA2 4,3 5,7 aB1. What are the NE
of this game?
Notice that Player B no longer has a strictly dominant action

Player B A1 is the prob. that A

aB1 aB2 chooses aA1.
B1 is the prob. that B
aA1 6,4 3,1 chooses aB1.
Player A
aA2 4,3 5,7
EVA(aA1) = ??
EVA(aA2) = ??

Player A
aA2 4,3 5,7
EVA(aA1) = 6B1 + 3(1 - B1) = 3 + 3B1.

EVA(aA2) = ??

Player A
aA2 4,3 5,7
EVA(aA1) = 6B1 + 3(1 - B1) = 3 + 3B1.

EVA(aA2) = 4B1 + 5(1 - B1) = 5 - B1.

Player A
aA2 4,3 5,7
EVA(aA1) = 6B1 + 3(1 - B1) = 3 + 3B1.

EVA(aA2) = 4B1 + 5(1 - B1) = 5 - B1.
3 + 3B1 >
= 5 -  B
1 as  B >
1 = ½.
< <
EVA(aA1) = 6B1 + 3(1 - B1) = 3 + 3B1.

EVA(aA2) = 4B1 + 5(1 - B1) = 5 - B1.
3 + 3B1 >
= 5 -  B
1 as  B >
1 = ½.
A ’ s best response < <
A1
1
0
0 ½ 1  B1
EVA(aA1) = 6B1 + 3(1 - B1) = 3 + 3B1.

EVA(aA2) = 4B1 + 5(1 - B1) = 5 - B1.
3 + 3B1 >
= 5 -  B
1 as  B >
1 = ½.
A ’ s best response < <
A1
1
0
0 ½ 1  B1

Player A
aA2 4,3 5,7
EVB(aB1) = ??
EVB(aB2) = ??

Player A
aA2 4,3 5,7
EVB(aB1) = 4A1 + 3(1 - A1) = 3 + A1.

EVB(aB2) = ??

Player A
aA2 4,3 5,7
EVB(aB1) = 4A1 + 3(1 - A1) = 4 + A1.

EVB(aB2) = A1 + 7(1 - A1) = 7 - 6A1.

Player A
aA2 4,3 5,7
EVB(aB1) = 4A1 + 3(1 - A1) = 3 + A1.

EVB(aB2) = A1 + 7(1 - A1) = 7 - 6A1.
4/
3 + A1 >
=7 - 6 A
1 as  A
1
>
=
7.
< <
EVB(aB1) = 4A1 + 3(1 - A1) = 3 + A1.

EVB(aB2) = A1 + 7(1 - A1) = 7 - 6A1.
4/
3 + A1 >
=7 - 6 A
1 as  A
1
>
=
7.
< <
A1
1
4/
7
0
0 1  B1
EVB(aB1) = 4A1 + 3(1 - A1) = 3 + A1.

EVB(aB2) = A1 + 7(1 - A1) = 7 - 6A1.
4/
3 + A1 >
=7 - 6 A
1 as  A
1
>
=
7.
< <
A1
1
4/
7
0
0 1  B1
B A
A ’ s best
A1 response
A1
1 1
4/
7
0 B
0
0 1  1 0 ½ 1  B1
B ’ s best
B A
A ’ s best
A1 response
A1
1 1
4/
7
0 B
0
0 1  1 0 ½ 1  B1
B A
A ’ s best
A1 response
A1
1 1
4/
7
0 B
0
0 1  1 0 ½ 1  B1
B ’ s best
A ’ s best
response
A1
1
4/
7
0
0 ½ 1  B1
B ’ s best
Is there a Nash equilibrium? Yes. 3 of them.
A ’ s best
response
A1
1
4/
7
0
0 ½ 1  B1
B ’ s best
Is there a Nash equilibrium? Yes. 3 of them.

(A1, B1) = (0,0)
A ’ s best
response
A1
1
4/
7
0
0 ½ 1  B1
B ’ s best
IsIsthere
thereaaNash
Nashequilibrium?
equilibrium? Yes. 3 of them.
(A1, B1) = (0,0)
(A1, B1) = (1,1)
A ’ s best
response
A1
1
4/
7
0
0 ½ 1  B1
B ’ s best
IsIsthere
thereaaNash
Nashequilibrium?
equilibrium? Yes. 3 of them.
(A1, B1) = (0,0)
(A1, B1) = (1,1)
A ’ s best 4/ ½
responseA1 (A1, B1) = ( 7, )
1
4/
7
0
0 ½ 1  B1
Some Important Types of Games
 Games of coordination
 Games of competition
 Games of coexistence
 Games of commitment
 Bargaining games
Coordination Games
 Simultaneous play games in which the
payoffs to the players are largest when
they coordinate their actions. Famous
examples are:
– The Battle of the Sexes Game
– The Prisoner’s Dilemma Game
– Assurance Games
– Chicken
Coordination Games; The Battle of the
Sexes
 Sissy prefers watching ballet to
watching mud wrestling.
 Jock prefers watching mud wrestling
to watching ballet.
 Both prefer watching something
together to being apart.
Sexes
Jock SB is the prob. that
B MW Sissy chooses ballet.
JB is the prob. that
8,4 1,2 Jock chooses ballet.
B
Sissy
MW 2,1 4,8
Sexes
B
What are the players’
Sissy best-response
functions?
MW 2,1 4,8
Sexes
B
Sissy best-response
functions?
MW 2,1 4,8
EVS(B) = 8JB + (1 - JB) = 1 + 7JB.

Sexes
B
Sissy best-response
functions?
MW 2,1 4,8
EVS(B) = 8JB + (1 - JB) = 1 + 7JB.

EVS(MW) = 2JB + 4(1 - JB) = 4 - 2JB.
Sexes
B
Sissy best-response
functions?
MW 2,1 4,8
EVS(B) = 8JB + (1 - JB) = 1 + 7JB.

EVS(MW) = 2JB + 4(1 - JB) = 4 - 2JB.
1 + 7JB >4 - 2JB as JB > 1/3.
=
< =
<
Sexes
B
SB
Sissy
1
MW 2,1 4,8
EVS(B) = 8JB + (1 - JB) = 1 + 7JB.

EVS(MW) = 2JB + 4(1 - JB) = 4 - 2JB.
1 + 7JB >4 - 2JB as JB > 1/3. 0
=
< =
< 0 1/3 1  JB
Sexes
B
SB Sissy
Sissy
1
MW 2,1 4,8
EVS(B) = 8JB + (1 - JB) = 1 + 7JB.

EVS(MW) = 2JB + 4(1 - JB) = 4 - 2JB.
1 + 7JB >4 - 2JB as JB > 1/3. 0
=
< =
< 0 1/3 1  JB
Sexes
SB Sissy S B Jock

1 1
1/3
0 0
0 1/3 J
1 B 0 1  JB
Sexes
The game’ s NE are ??

1 1
1/3
0 0
0 1/3 J
1 B 0 1  JB
Sexes

1 1
1/3
0 0
0 1/3 J
1 B 0 1  JB
Sexes
Sissy
SB
1
1/3
0
0 1/3 1  JB
Jock
Sexes
The game’ s NE are: (JB, SB) = (0, 0); i.e., (MW, MW)
Sissy
SB
1
1/3
0
0 1/3 1  JB
Jock
Sexes
(JB, SB) = (1, 1); i.e., (B, B)
Sissy
SB
1
1/3
0
0 1/3 1  JB
Jock
Sexes
(JB, SB) = (1, 1); i.e., (B, B)
( J
, S
) = ( 1,/3 1);
/
3 i.e., both
Sissy B  B
SBwatch the ballet with prob. 1/9, both watch the mud
wrestling with prob. 4/9, and with prob. 4/9 they
1
watch different events.
1/3
0
0 1/3 1  JB
Jock
Sexes
B
Sissy
MW 2,1 4,8
or Sissy the expected value of the NE (JB, SB) =1/(3 1/,3 ) is

× 1/9 + 1× 2/9+ 2× 2/9 + 4× 4/9 =10/3 < 4 and 8.
Sexes
B
Sissy
MW 2,1 4,8

× 1/9 + 1× 2/9+ 2× 2/9 + 4× 4/9 =10/3 < 4 and 8.
or Jock the expected value of the NE (JB, SB) =1(/3 1/,3 ) is
4× 1/9+ 2× 2/9+ 1× 2/9 + 8×4/9 =14/3 ; 4 < 14/3 < 8.
Sexes
B
So, is the mixed
Sissy
strategy NE a focal
MW 2,1 4,8 point for the game?

× 1/9 + 1× 2/9+ 2× 2/9 + 4× 4/9 =10/3 < 4 and 8.
or Jock the expected value of the NE (JB, SB) =1(/3 1/,3 ) is
4× 1/9+ 2× 2/9+ 1× 2/9 + 8×4/9 =14/3 ; 4 < 14/3 < 8.
Coordination Games; The Prisoner’s
Dilemma
 A simultaneous play game in which
each player has a strictly dominant
action.
Dilemma
 The only NE, therefore, is the choice
by each player of her strictly
dominant action.
 Yet both players can achieve strictly
larger payoffs than in the NE by
coordinating with each other on
another pair of actions.
Dilemma
 Tim and Tom are in police custody.
Each can confess (C) to a crime or
stay silent (S).
 Confession by both results in 5 years
each in jail.
Dilemma
 Silence by both results in 2 years
each in jail.
 If Tim confesses and Tom stays
silent then Tim gets no penalty and
Tom gets 10 years in jail (and
conversely).
Dilemma
Tom
Silent Confess
Silent -2,-2 -10,0

Tim
Confess 0,-10 -5,-5
For Tim, Confess strictly dominates Silent.

Dilemma
Tom
Silent Confess
Silent -2,-2 -10,0

Tim
Confess 0,-10 -5,-5

For Tom, Confess strictly dominates Silent.
Dilemma
Tom
Silent Confess
Silent -2,-2 -10,0

Tim
Confess 0,-10 -5,-5

The only NE is (Confess, Confess).
Dilemma
Tom
Silent Confess
But (Silence, Silence)
is better for both Tim
Silent -2,-2 -10,0
and Tom.
Tim
Confess 0,-10 -5,-5

The only NE is (Confess, Confess).
Dilemma
Tom What is needed is a
Silent Confess means of rationally
assuring commitment
by both players to
Silent -2,-2 -10,0 the most beneficial
Tim coordinated actions.
Confess 0,-10 -5,-5
Possible means include future punishments or enforceable

contracts.
Coordination Games; Assurance Games
 A simultaneous play game with two

“coordinated” NE, one of which
gives strictly greater payoffs to each
player than does the other.
 The question is: How can each
player give the other an “assurance”
that will cause the better NE to be the
outcome of the game?
 A common example is the “arms rac

e” problem.
 India and Pakistan can both increase
their stockpiles of nuclear weapons.
This is very costly.
 Having nuclear superiority over the

other gives a higher payoff, but the
worst payoff to the other.
 Not increasing the stockpile is best
for both.
Pakistan
Don’ t Stockpile
Don’ t 5,5 1,4

India
Stockpile 4,1 3,3
Pakistan
Don’ t Stockpile
Don’ t 5,5 1,4

India
Stockpile 4,1 3,3

Pakistan
Don’ t Stockpile
Don’ t 5,5 1,4

India
Stockpile 4,1 3,3
e game’ s NE are (Don’ t, Don’ t) and (Stockpile, Stockpile)

ich is the “ likely” NE?
Pakistan
Don’ t Stockpile
Don’ t 5,5 1,4

India
Stockpile 4,1 3,3
e game’ s NE are (Don’ t, Don’ t) and (Stockpile, Stockpile)

ich is the “ likely” NE? What if India moved first? What
ion would it choose? Wouldn’ t Don’ t be best?
Coordination Games; Chicken
 A simultaneous play game with two

“coordinated” NE in which each
player chooses the action that is not
the action chosen by the other
player.
 Two drivers race their cars at each

other. A driver who swerves is a
“wimp”. A driver who does not
swerve is “macho.”
 If both do not swerve there is a crash
and a very low payoff to both.
 If both swerve then there is no crash

and a moderate payoff to both.
 If one swerves and the other does
not then the swerver gets a low
payoff and the non-swerver gets a
high payoff.
Dumber No
Swerve Swerve
Swerve 1,1 -2,4

Dumb
No Swerve 4,-2 -5,-5

Dumber No
Swerve Swerve
Swerve 1,1 -2,4

Dumb
No Swerve 4,-2 -5,-5
he game’ s pure strategy NE are (Swerve, No Swerve) and

No Swerve, Swerve). There is also a mixed strategy NE in
which each chooses Swerve with probability ½.
Dumber No Can Dumb assure

Swerve Swerve himself of a payoff of
4? Only by convincing
Swerve 1,1 -2,4 Dumber that Dumb
really will choose No
Dumb Swerve. What will be
No Swerve 4,-2 -5,-5 convincing?
he game’ s pure strategy NE are (Swerve, No Swerve) and

No Swerve, Swerve). There is also a mixed strategy NE in
which each chooses Swerve with probability ½.
Games of Competition
 Simultaneous play games in which

any increase in the payoff to one
player is exactly the decrease in the
payoff to the other player.
 These games are thus often called
“constant (payoff) sum” games.
An example is the game below. What NE can such a

game possess?
2
L R
U 0,0 2,-2
1
D x,-x 1,-1

game possess?
If x < 0 then Up ??
2
L R
U 0,0 2,-2
1
D x,-x 1,-1

game possess?
If x < 0 then Up dominates
Down.
2
L R
U 0,0 2,-2
1
D x,-x 1,-1

game possess?
Down.
2 If x < 1 then Left ??
L R
U 0,0 2,-2
1
D x,-x 1,-1

game possess?
Down.
2 If x < 1 then Left dominates
L R Right.
U 0,0 2,-2
1
D x,-x 1,-1

game possess?
Down.
L R Right.
Therefore, if x < 0 the NE
is ??
U 0,0 2,-2
1
D x,-x 1,-1

game possess?
Down.
L R Right.
is (Up, Left)
U 0,0 2,-2
1
D x,-x 1,-1

game possess?
Down.
L R Right.
is (Up, Left) and if 0 < x < 1
U 0,0 2,-2
the NE is ??
1
D x,-x 1,-1

game possess?
Down.
L R Right.
U 0,0 2,-2
the NE is (Down, Left).
1
D x,-x 1,-1

game possess?
Down.
L R Right.
U 0,0 2,-2
1 If x > 1 then ??
D x,-x 1,-1

game possess?
Down.
L R Right.
U 0,0 2,-2
1 If x > 1 then there is no NE
in pure strategies. Is there
D x,-x 1,-1
a mixed-strategy NE?
The probability that 2 chooses

Left is L. The probability that
1 chooses Up is U. x > 1.
2
L R
U 0,0 2,-2
1
D x,-x 1,-1
EV1(U) = 2(1 - L).

Left is L. The probability that EV1(D) = xL + 1 - L.
1 chooses Up is U. x > 1.
2
L R
U 0,0 2,-2
1
D x,-x 1,-1
EV1(U) = 2(1 - L).

2 - 2l 1 + (x - 1)L
1 chooses Up is U. x > 1. >
=
2 as L <1/(1 + x).
L R >
=
<
U 0,0 2,-2
1
D x,-x 1,-1
EV1(U) = 2(1 - L).

2 - 2l 1 + (x - 1)L
=
2 as L <1/(1 + x).
L R >
=
<
EV2(L) = - x(1 - U).
U 0,0 2,-2 EV2(R) = - 2U - (1 - U).
1
D x,-x 1,-1
EV1(U) = 2(1 - L).

2 - 2l 1 + (x - 1)L
=
2 as L <1/(1 + x).
L R >
=
<
EV2(L) = - x(1 - U).
U 0,0 2,-2 EV2(R) = - 2U - (1 - U).
1 - x + xU - 1 - U
>
=
D x,-x 1,-1 < + x)
as (x – 1)/(1 U.
>
=
<
1 chooses Up if L > 1/(1 + x) and Down if L < 1/(1 + x).

2 chooses Left if U < (x – 1)/(1 + x) and
Right if U > (x – 1)/(1 + x).
2
L R
U 0,0 2,-2
1
D x,-x 1,-1
1: Up if L > 1/(1 + x); Down if L < 1/(1 + x).

2: Left if U < (x – 1)/(1 + x); Right if U > (x – 1)/(1 + x).
U 1 L 2
1 1
0 0
0 L 0 1 U
1: Up if L > 1/(1 + x); Down if L < 1/(1 + x).

U 1 L 2
1 1
0 0
0 1/(1+x) L 0 (x-1)/(1+x) 1 U
1: Up if L > 1/(1 + x); Down if L < 1/(1 + x).

U 1 L 2
1 1
0 0
0 1/(1+x) L 0 (x-1)/(1+x) 1 U
1: Up if L > 1/(1 + x); Down if L < 1/(1 + x).

U 1 L 2
1 1
0 0
0 1/(1+x) L 0 (x-1)/(1+x) 1 U
1: Up if L > 1/(1 + x); Down if L < 1/(1 + x).

U 1 L 2
1 1
(x-1)/(1+x)
0 0
0 1/(1+x) L 0 (x-1)/(1+x) 1 U
1: Up if L > 1/(1 + x); Down if L < 1/(1 + x).

U 1 U 2
1 1
(x-1)/(1+x)
0 0
0 1/(1+x) L 0 1 L
1: Up if L > 1/(1 + x); Down if L < 1/(1 + x).

U 1 U 2
1 1
(x-1)/(1+x)
0 0
0 1/(1+x) L 0 1 L
1: Up if L > 1/(1 + x); Down if L < 1/(1 + x).

U
1
When x > 1 there is only
a mixed-strategy NE in
which 1 plays Up with (x-1)/(1+x)
probability (x – 1)/(x + 1)
and 2 plays Left with
probability 1/(1 + x).
0
0 1/(1+x) 1 L
Coexistence Games
 Simultaneous play games that can be

used to model how members of a
species act towards each other.
 An important example is the hawk-
dove game.
Coexistence Games; The Hawk-Dove
Game
 “Hawk” means “be aggressive.”
 “Dove” means “don’t be aggressive.”
 Two bears come to a fishing spot.
Either bear can fight the other to try to
drive it away to get more fish for itself
but suffer battle injuries, or it can
tolerate the presence of the other,
share the fishing, and avoid injury.
Game
Bear 2
Hawk Dove
Hawk -5,-5 8,0

Bear 1
Dove 0,8 4,4
Are there NE in pure strategies?

Game
Bear 2
Hawk Dove
Hawk -5,-5 8,0

Bear 1
Dove 0,8 4,4
Are there NE in pure strategies?

Yes (Hawk, Dove) and (Dove, Hawk).
Notice that purely peaceful coexistence is not a NE.
Game
Bear 2
Hawk Dove
Hawk -5,-5 8,0

Bear 1
Dove 0,8 4,4
Is there a NE in mixed strategies?

Game
Bear 2 1H is the prob. that
Hawk Dove 1 chooses Hawk.
2H is the prob. that
-5,-5 8,0 2 chooses Hawk.
Hawk
Bear 1 best-response
functions?
Dove 0,8 4,4
Is there a NE in mixed strategies?

Game
Hawk
functions?
Dove 0,8 4,4
EV1(H) = -52H + 8(1 - 2H) = 8 - 132H.

EV1(D) = 4 - 42H.
Game
Hawk
functions?
Dove 0,8 4,4
EV1(H) = -52H + 8(1 - 2H) = 8 - 132H.

EV1(D) = 4 - 42H.
8 - 132H >
=4 - 4 2
H as  2 <
H = 4/9.
< >
Game
Hawk 1H Bear 1
Bear 1
1
Dove 0,8 4,4
EV1(H) = -52H + 8(1 - 2H) = 8 - 132H.

EV1(D) = 4 - 42H.
8 - 132H >
=4 - 4 2
H as  2 <
H = 4/9. 0
< > 0 4/9 1 2H
Game
1H Bear 1 2H Bear 2
1 1
0 2
0
0 4/9 1  H 0 4/9 1 1H
Game
1 1
4/9
0 2
0
0 4/9 1  H 0 4/9 1 1H
Game
1 1
4/9
0 2
0
0 4/9 1  H 0 1 2H
Game
The game has a NE in mixed-strategies in which

each bear plays Hawk with probability 4/9.
1H
0
0 4/9 1 2H
Game
Bear 2
Hawk Dove
Hawk -5,-5 8,0

Bear 1
Dove 0,8 4,4
r each bear, the expected value of the mixed-strategy NE is

5)× 8× 20/81
16+/81 + 4× 25/81= 18081 / , a value between
and +4. Is this NE focal?
Commitment Games
 Sequential play games in which
– One player chooses an action before
the other player chooses an action.
– The first player’s action is both
irreversible and observable by the
second player.
– The first player knows that his action
is seen by the second player.
Commitment Games
5,9
c Player 1 has two
Game Tree 2 actions, a and b.
Player 2 has two
a actions, c and d,
d 5,5 following a, and
1 two actions e and
Direction of play
f
7,6
e following b.
b
Player 1 chooses
2 his action before
f 5,4 Player 2 chooses
Commitment Games
5,9 Is a claim by Player 2

c that she will commit to
2 choosing action c if
a Player 1 chooses a
credible to Player 1?
d 5,5
1
7,6
e
b
2
f 5,4
Commitment Games

d 5,5 Yes.
1
7,6
e
b
2
f 5,4
Commitment Games

d 5,5 Yes.
1
e
that she will commit to
b
choosing action e if
2 Player 1 chooses b
f 5,4
Commitment Games

d 5,5 Yes.
1
e
that she will commit to
b
choosing action e if
2 Player 1 chooses b
f 5,4 Yes.
Commitment Games
5,9 So Player 1 should

c choose action ??
2
a
d 5,5
1
7,6
e
b
2
f 5,4
Commitment Games
5,9 So Player 1 should

c choose action b.
2
a
d 5,5
1
7,6
e
b
2
f 5,4
Commitment Games
5,3 Change the game.

c
2
a
d 5,5
1
7,6
e
b
2
f 5,4
Commitment Games

d 5,5
1
7,6
e
b
2
f 5,4
Commitment Games

d 5,5
1 No. If Player 1 chooses
7,6 action a then Player 2
e
does best by choosing
b
action d.
2 What should Player 1
do?
f 5,4
Commitment Games

d 5,5
1 No. If Player 1 chooses
7,6 action a then Player 2
e
does best by choosing
b
action d.
2 What should Player 1
do? Still choose b.
f 5,4
Commitment Games
5,3 Change the game.

5,9
c
2
a
d 15,5
5,5
1
7,3
e
b
2
f 5,12
Commitment Games
5,9 Can Player 1 get 15

c points?
2
a
d 15,5
1
7,3
e
b
2
f 5,12
Commitment Games

c points?
2
a If Player 1 chooses a
then Player 2 will
d 15,5 choose c and Player 1
1 will get only 5 points.
7,3
e
b
2
f 5,12
Commitment Games

c points?
2
a If Player 1 chooses a
then Player 2 will
d 15,5 choose c and Player 1
1 will get only 5 points.
7,3
e
If Player 1 chooses b
b
then Player 2 will
2 choose f and again
Player 1 will get only
f 5,125 points.
Commitment Games
5,9 If Player 1 can change

c payoffs so that a
2 commitment by Player
2 to choose d after a
a is credible then Player
d 15,5 1 ’ s payoff rises from 5
1 to 15, a gain of 10.
7,3
e
b
2
f 5,12
Commitment Games
5,9 If Player 1 can change

c payoffs so that a
2 commitment by Player
2 to choose d after a
a is credible then Player
d 1 ’ s payoff rises from 5
10,10
1 to 15, a gain of 10.
7,3 If Player 1 gives 5 of
e these points to Player
b 2 then Player 2’ s
commitment is
2
credible. Player 1
f 5,12cannot get 15 points.
Commitment Games
5,9 Credible NE of this

c type are called
2 subgame perfect.
What exactly is this
a game’ s SPE? It
d 10,10insists that every
1 action chosen is
7,3 rational for the player
e who chooses it.
b
2
f 5,12
Commitment Games

c type are called
2 subgame perfect.
a game’ s SPE? It
1 action chosen is
e who chooses it.
b
2
f 5,12
Commitment Games

c type are called
2 subgame perfect.
a game’ s SPE? It
1 action chosen is
e who chooses it.
b
2
f 5,12
Commitment Games

c type are called
2 subgame perfect.
a game’ s SPE? It
1 action chosen is
e who chooses it.
b
2
f 5,12
Commitment Games

c type are called
2 subgame perfect.
a game’ s SPE? It
1 action chosen is
e who chooses it.
b
2
f 5,12
Bargaining Games
 Two players bargain over the division

of a pie of size 1. What will be the
outcome?
 Two approaches:
– Nash’s axiomatic bargaining.
– Rubinstein’s strategic bargaining.
Strategic Bargaining
 The players have 3 periods in which to decide
how to divide the pie; else both get nothing.
 Player A discounts next period’s payoffs by .
 Player B discounts next period’s payoffs by .
 The players alternate in making offers, with
Player A starting in period 1.
 If the player who receives an offer accepts it
then the game ends immediately. Else the
game continues to the next period.
Period 2:
(x1,1-x1) B offers x2. (x3,1-x3)
A responds.
1 Y 1 1 Y
x1 x3
N
B B (0,0)
N
A B x2 A A
N
Y
0 0 0
Period 1: Period 3:
A offers x1. (x2,1-x2) A offers x3.
B responds. B responds.
How should B respond to x3?
(x3,1-x3)
1 Y
x3
N
B (0,0)
A
0
Period 3:
A offers x3.
B responds.
Accept if 1 – x3 ≥ 0; i.e., (x3,1-x3)
accept any x3 ≤ 1.
1 Y
x3
N
B (0,0)
A
0
Period 3:
A offers x3.
B responds.
Accept if 1 – x3 ≥ 0; i.e., (x3,1-x3)
accept any x3 ≤ 1.
1 Y
Knowing this, what should A
offer? x3
N
B (0,0)
A
0
Period 3:
A offers x3.
B responds.
Accept if 1 – x3 ≥ 0; i.e., (1,0)
accept any x3 ≤ 1. Y
x3=1
B
Knowing this, what should A
offer? x3 = 1. N
(0,0)
A
0
Period 3:
A offers x3 = 1.
B accepts.
Period 2:
(x1,1-x1) B offers x2. (1,0)
A responds. Y
1 Y 1 x3=1
B
x1
N
B N (0,0)
A B x2 A A
N
Y
0 0 0
Period 1: Period 3:
A offers x1. (x2,1-x2) A offers x3 = 1.
B responds. B accepts.
In Period 3 A gets Period 2:

a payoff of 1. In B offers x2. (1,0)
period 2, when A responds. Y
replying to B’ s 1 x3=1
B
offer of x2, the
present-value to N
A of N is thus ?? (0,0)
B x2 A A
N
Y
0 0
Period 3:
(x2,1-x2) A offers x3 = 1.
B accepts.

a payoff of 1. In B offers x2. (1,0)
period 2, when A responds. Y
replying to B’ s 1 x3=1
B
offer of x2, the
present-value to N
A of N is thus . (0,0)
B x2 A A
N
Y
0 0
Period 3:
(x2,1-x2) A offers x3 = 1.
B accepts.

a payoff of 1. In B offers x2.
period 2, when A responds.
replying to B’ s 1
offer of x2, the
present-value to
A of N is thus .
B x2 A
What is the most N
B should offer to Y
A? 0
(x2,1-x2)

a payoff of 1. In B offers x2 = .
period 2, when A accepts.
replying to B’ s 1
offer of x2, the
present-value to
A of N is thus . x2=
B A
What is the most N
B should offer to Y
A? x2 = . 0
(,1- )
Period 2:
(x1,1-x1) B offers x2 = . (1,0)
A accepts. Y
1 Y 1 x3=1
B
x1
N
B N x2= (0,0)
A B A A
N
Y
0 0 0
Period 1: Period 3:
A offers x1. (,1- ) A offers x3 = 1.
B responds. B accepts.
Period 2: In period 2 A will accept
(x1,1-x1) B offers x2 = .. Thus B will get the
A accepts. payoff 1 -  in period 2.
1 1 What is the present-
Y
value to B in period 1 of
x1 N?
B N x2=
A B A
N
Y
0 0
Period 1:
A offers x1. (,1- )
B responds.
Y
x1 N ? (1 - ).
B N x2=
A B A
N
Y
0 0
Period 1:
A offers x1. (,1- )
B responds.
Y
x1 N ? (1 - ).
B N x2= What is the most that A
A B A
N should offer to B in
period 1?
Y
0 0
Period 1:
A offers x1. (,1- )
B responds.
(1-(1 - ), Period 2: In period 2 A will accept
(1 - )) B offers x2 = .. Thus B will get the
Y
x1 N ? (1 - ).
B N x2= What is the most that A
A B A
N should offer to B in
period 1?
Y 1 – x1 = (1 - ); i.e.
0 0
Period 1: x1 = 1 - (1 - ).
A offers x1. (,1- ) B will accept.
B responds.
(1-(1 - ), Period 2:
(1 - )) B offers x2 = . (1,0)
A accepts. Y
x1=1- 1 Y 1 x3=1
B
(1-)
N
B N x2= (0,0)
A B A A
N
Y
0 0 0
Period 1: A offers Period 3:
x1 = 1-(1 - ). (,1- ) A offers x3 = 1.
B accepts. B accepts.
 Notice that the game ends

immediately, in period 1.
 Player A gets 1 - (1 – ) units of the

pie.
Player B gets (1 – ) units.
Which is the larger?
 x1 = 1 - (1 – ) ≥ ½   ≤ 1/2(1 - )
so Player A gets more than Player B
if Player B is “too impatient” relative
to Player A.
 Suppose the game is allowed to
continue forever (infinitely many
periods). Then using the same
reasoning shows that the subgame
perfect equilibrium results in
Players 1 and 2 respectively getting
1   (1   )
and pie units.
1   1  
 Player 1’s share rises as  and  .
Player 2’s share rises as  and  .

Varian9e LecturePPTs Ch30

Uploaded by

Document Information

Original Description:

Original Title

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

Varian9e LecturePPTs Ch30

Uploaded by

Copyright:

Available Formats

Chapter 30

 In any Nash equilibrium (NE) each

 How can we locate every one of a

 Think of a 2×2 game; i.e., a game

 B can choose between actions aB1

 UA(aA1, aB1) = 6 and UB(aA1, aB1) = 4

 UA(aA1, aB1) = 6 and UB(aA1, aB1) = 4

 UA(aA1, aB1) = 6 and UB(aA1, aB1) = 4

 UA(aA1, aB1) = 6 and UB(aA1, aB1) = 4

 UA(aA1, aB1) = 6 and UB(aA1, aB1) = 4

 UA(aA1, aB1) = 6 and UB(aA1, aB1) = 4

 UA(aA1, aB1) = 6 and UB(aA1, aB1) = 4

 UA(aA1, aB1) = 6 and UB(aA1, aB1) = 4

 UA(aA1, aB1) = 6 and UB(aA1, aB1) = 4

 UA(aA1, aB1) = 6 and UB(aA1, aB1) = 4

How can the players’ best-response curves be

How can the players’ best-response curves be

How can the players’ best-response curves be

How can the players’ best-response curves be

How can the players’ best-response curves be

How can the players’ best-response curves be

Player B Here is the strategic

aA1 6,4 3,5

aA2 is the only best response to aB2.

Player B Here is the strategic

aA2 is the only best response to aB2.

Player B Here is the strategic

aA2 is the only best response to aB2.

aA1 6,4 3,5

Now allow both players to randomize (i.e., mix)

Now allow both players to randomize (i.e., mix)

A1 is the prob. A chooses action aA1.

A1 is the prob. A chooses action aA1.

A1 is the prob. A chooses action aA1.

A1 is the prob. A chooses action aA1.

A ’ s best response is: aA1 (i.e. A1 = 1) if B1

A ’ s best response is: aA1 (i.e. A1 = 1) if

A ’ s best response is: aA1 (i.e. A1 = 1) if B1

A ’ s best response is: aA1 (i.e. A1 = 1) if B1

A1 is the prob. A chooses action aA1.

A1 is the prob. A chooses action aA1.

A1 is the prob. A chooses action aA1.

B ’ s best response is aB2 always (i.e. B1 = 0

Is there a Nash equilibrium?

Is there a Nash equilibrium?

Is there a Nash equilibrium?

Is there a Nash equilibrium? Yes. Just one.

Player B Let’ s change the game

aA1 6,4 3,5

Player B Here is a new

aA1 6,4 3,1

Player B Here is a new

Notice that Player B no longer has a strictly dominant action

Player B A1 is the prob. that A

Player B A1 is the prob. that A

EVA(aA1) = 6B1 + 3(1 - B1) = 3 + 3B1.

Player B A1 is the prob. that A

EVA(aA1) = 6B1 + 3(1 - B1) = 3 + 3B1.

Player B A1 is the prob. that A

EVA(aA1) = 6B1 + 3(1 - B1) = 3 + 3B1.

EVA(aA1) = 6B1 + 3(1 - B1) = 3 + 3B1.

EVA(aA1) = 6B1 + 3(1 - B1) = 3 + 3B1.