ec941 - game theory
DESCRIPTION
EC941 - Game Theory. Lecture 2. Prof. Francesco Squintani Email: [email protected]. Structure of the Lecture. Mixed Strategies Nash Equilibrium and Rationalizability Correlated Equilibrium. Mixed Strategies. In a game in strategic form G=(I, S, u), for each player i , S i - PowerPoint PPT PresentationTRANSCRIPT
![Page 2: EC941 - Game Theory](https://reader034.vdocuments.net/reader034/viewer/2022051219/56816020550346895dcf2121/html5/thumbnails/2.jpg)
Structure of the Lecture
Mixed Strategies
Nash Equilibrium and Rationalizability
Correlated Equilibrium
2
![Page 3: EC941 - Game Theory](https://reader034.vdocuments.net/reader034/viewer/2022051219/56816020550346895dcf2121/html5/thumbnails/3.jpg)
Mixed Strategies
In a game in strategic form G=(I, S, u), for each player i, Si
is the set of pure strategies.
A mixed strategy si is a probability distribution over Si.
When playing si , player i operates a randomizing device and
chooses the strategy accordingly.
For example, she flips a coin choosing one strategy if heads
turns up, and another one if tails turns up.
3
![Page 4: EC941 - Game Theory](https://reader034.vdocuments.net/reader034/viewer/2022051219/56816020550346895dcf2121/html5/thumbnails/4.jpg)
In general, randomizing devices can be very complex and
induce any probability distribution.
Payoffs and best-response correspondences can be extended
from pure strategies to mixed strategies.
Solution concepts (Nash Equilibrium, Dominance and
Rationalizability) can be extended to mixed strategies.
To play a mixed strategy si, player i must be indifferent
among the possible pure strategies.
This indifference principle is crucial to calculate solutions.
4
![Page 5: EC941 - Game Theory](https://reader034.vdocuments.net/reader034/viewer/2022051219/56816020550346895dcf2121/html5/thumbnails/5.jpg)
Example: Matching Pennies
HeadTail
Head
Tail
-1, 1
1, -1
A reasonable solution is that both players randomize between H and T with equal probability. The game results in a tie.
-1, 1
1, -1 There is no sure way to win for either of the players.
5
![Page 6: EC941 - Game Theory](https://reader034.vdocuments.net/reader034/viewer/2022051219/56816020550346895dcf2121/html5/thumbnails/6.jpg)
Let p, 0<p<1, be player 1’s probability to play H, and q, 0<q<1, be player 2’s probability to play H.
Player 1’s expected payoff for playing H is: U1 = -1 . q + 1 . (1-q).
Player 1’s payoff for T is U1 = 1 . q - 1 . (1-q).
1 q0
1
-1
H
T
U1 When player 2 mixes between T and H with equal probability, player 1 is indifferent between T and H.
6
![Page 7: EC941 - Game Theory](https://reader034.vdocuments.net/reader034/viewer/2022051219/56816020550346895dcf2121/html5/thumbnails/7.jpg)
We can write best response correspondences in terms of mixed strategy:B1(q) = 1 if q < ½ B2(p) = 0 if p < ½B1(q) = [0,1] if q = ½ B2(p) = [0,1] if p = ½ B1(q) = 0 if q > ½ B2(p) = 1 if p > ½.
0 ½ 1 p
q1
½
The Nash Equilibrium is found by intersecting theBest Response correspondences.
The Nash Equilibrium is (p, q) = (½, ½).
B1(q)
B2(p)
7
![Page 8: EC941 - Game Theory](https://reader034.vdocuments.net/reader034/viewer/2022051219/56816020550346895dcf2121/html5/thumbnails/8.jpg)
General Definitions
Consider a game in strategic form G = (I, S, u).
Definition A mixed strategy si of a player i is a probability
distribution over player i’s pure strategies Si.
The set of all player i’s possible mixed strategies is
Si ={(si (si)) : si (si) > 0 for all si, and Ss si (si) = 1}.
8
![Page 9: EC941 - Game Theory](https://reader034.vdocuments.net/reader034/viewer/2022051219/56816020550346895dcf2121/html5/thumbnails/9.jpg)
Definition When the players play the mixed strategy
profile s, player i’s expected payoff Ui is:
Ui (s) = Ss [ Pi si (si) ] ui (s).
Definition The mixed strategy extension of G is a game
in strategic form G = (I, S, U).
Note The mixed strategies S of the game G are the pure
strategies of game G.
Best Response Correspondences, Nash Equilibrium,
Dominance and Rationalizability can be defined for game G.
9
![Page 10: EC941 - Game Theory](https://reader034.vdocuments.net/reader034/viewer/2022051219/56816020550346895dcf2121/html5/thumbnails/10.jpg)
The mixed strategy profile s∗ is a Nash equilibrium if,
for each player i, Ui(s∗) ≥ Ui(si, s∗−i) for
every mixed strategy si of player i.
For every player i, the best response correspondence isBi (s−i) = {si : Ui(si, s−i) ≥ Ui(s’i, s−i) for
every s’i}.
The profile s∗ is a Nash equilibrium if and only if s∗
i belongs to Bi (s∗−i), for every i.
Player i’s mixed strategy si strictly dominates her strategy si if Ui(si, s−i) > ui(si, s−i) for every s−i.
10
![Page 11: EC941 - Game Theory](https://reader034.vdocuments.net/reader034/viewer/2022051219/56816020550346895dcf2121/html5/thumbnails/11.jpg)
Proposition If player i’s strategy si is strictly dominated, then any
mixed strategy si which assigns strictly positive probability to si is also
strictly dominated.
The definition of rationalizability follows from the
iterated deletion of strictly dominated strategies.
Given a game G = (I, S, U), let Xi1 = Si.
For each t = 0, . . . , T − 1, Xit+1 is a subset of
Xit such that
every si in Xit that is not in Xi
t+1 is strictly dominated in
G = (I, (Xit)I, U).
The set XiT is the set of rationalizable
strategies of player i.
11
![Page 12: EC941 - Game Theory](https://reader034.vdocuments.net/reader034/viewer/2022051219/56816020550346895dcf2121/html5/thumbnails/12.jpg)
The Indifference Principle
Indifference Principle. If the opponents play s−i, player i is
indifferent among all pure strategies in the best response set Bi (s−i).
Definition For any si, let Ci(si) be the support of si, the
set of strategies si to which si assigns non-zero probability.
Proposition The profile s* is a Nash Equilibrium if and only if
Ci(s*i) is a subset of Bi(s*
−i), for every player i.
12
![Page 13: EC941 - Game Theory](https://reader034.vdocuments.net/reader034/viewer/2022051219/56816020550346895dcf2121/html5/thumbnails/13.jpg)
13
In games with 2 players and 2 strategies for each player
(let A1 = {U, D} and A2 = {L, R}), we find thecompletely mixed strategy Nash equilibrium
as follows.
From the indifference condition for player 1: sL u1(U,L) + (1- sL) u1(U,R) = sL u1(D,L) + (1- sL) u1(D,R)
we find the mixed strategy sL.
sU u2(U,L) + (1-sU) u2(D,L) = sU u2(U,R) + (1-sU) u2(D,R)
pins down the mixed strategy sU.
If 0<sL<1 and 0<sU<1, then (sL, sU) is Nash Equilibrium.
![Page 14: EC941 - Game Theory](https://reader034.vdocuments.net/reader034/viewer/2022051219/56816020550346895dcf2121/html5/thumbnails/14.jpg)
Definition For any player i, and (non-empty) subset Bi of
the strategy set Si, let S−i (Bi) be the set of mixed strategy
profiles s−i such that Bi = Bi(s−i).
We can find all mixed strategy equilibria as follows.
Consider each profile B of subsets Bi, and calculate the
profile (S−i (Bi))I. The mixed strategy s is a Nash Equilibrium
if:1. s−i belongs to S−i (Bi) for all players i. 2. Ci(si) = Bi for all players i.
14
![Page 15: EC941 - Game Theory](https://reader034.vdocuments.net/reader034/viewer/2022051219/56816020550346895dcf2121/html5/thumbnails/15.jpg)
Let p, q be player 1’s and 2’s probability to play B.
U1(B) = 2・ q + 0・ (1 − q) = 2q. U1(S) = 0・ q + 1・ (1 − q) = 1 − q. The condition U1(B) = U1(S), i.e. 2q = 1 − q
yields q = 1/3.
Battle of the SexesB S
B
S
2, 1
0, 0 1, 2
0, 0
1 q
1
2
0
B
S
U1
1/3
15
![Page 16: EC941 - Game Theory](https://reader034.vdocuments.net/reader034/viewer/2022051219/56816020550346895dcf2121/html5/thumbnails/16.jpg)
U2(B) = 1・ p + 0・ (1 − p) = p. U2(S) = 0・ p + 2・ (1 − p) = 2 − 2p. The condition U2(B) = U2(S), yields p = 2/3.
B1 (q) = 0 if q < 1/3B1 (q) = [0, 1] if q = 1/3B1 (q) =1 if q > 1/3
B2 (p) = 0 if p < 2/3B2 (p) = [0, 1] if p = 2/3B2 (p) =1 if p > 2/3
q 1
1/3
B1(q)B2(p)
0 2/3 1 p16
![Page 17: EC941 - Game Theory](https://reader034.vdocuments.net/reader034/viewer/2022051219/56816020550346895dcf2121/html5/thumbnails/17.jpg)
The game has three mixed strategy Nash equilibria: (p, q) = (0, 0), (2/3, 1/3), and (1, 1).
The mixed strategy equilibria (0, 0) and (1, 1) correspond to the two pure strategy equilibria.
The expected payoff of the equilibrium (2/3, 1/3) is U1 = 2 ・ 2/9 + 0・ 4/9 + 0・ 1/9 + 1・ 2/9 = 2/3.Likewise, U2 = 2/3.
So, it is Pareto dominated by both pure strategy equilibria, which yield payoffs (2, 1) and (1, 2).
17
![Page 18: EC941 - Game Theory](https://reader034.vdocuments.net/reader034/viewer/2022051219/56816020550346895dcf2121/html5/thumbnails/18.jpg)
Let p, q be player 1’s and 2’s probability to play E.
U1(N) = 0, U1(E) = (1 – c)q - c(1 − q) = q - c.
The condition U1(N) = U1(E), yields q = c.
The case of player 2 is symmetric.
N E
N
E
0, 0
-c, 0 1-c, 1-c
0, -c1 q
0
1-c
-c
EN
U1
c
Coordination Game
18
![Page 19: EC941 - Game Theory](https://reader034.vdocuments.net/reader034/viewer/2022051219/56816020550346895dcf2121/html5/thumbnails/19.jpg)
B1(q) = 0 if q < cB1(q) = [0, 1] if q = cB1(q) =1 if q > c
B2(p) = 0 if p < cB2(p) = [0, 1] if p = cB2(p) =1 if p > c
The game has three mixed strategy Nash equilibria:
(p, q) = (0, 0), (c, c), and (1, 1).The mixed strategy equilibria (0, 0) and (1,
1) correspondto the pure strategy equilibria (N, N) and
(E, E).The expected payoff of the equilibrium (c, c)
is U1 = c2(1-c) - c(1-c)c + (1-c)0 = 0 = U2.
B1(q)B2(p)
0 c 1 p
19
![Page 20: EC941 - Game Theory](https://reader034.vdocuments.net/reader034/viewer/2022051219/56816020550346895dcf2121/html5/thumbnails/20.jpg)
Games with Dominated Strategies
A B C
2, 1
1, 3
1, 1
0, 2
3, 1
1, 1
4, 1
3, 0
2, 2
D E F
Strategy C is dominated by any mixed strategy a1 with 0 < a1(A) < 2/3. Once C is deleted, strategy F is dominated by E.
1
2
20
![Page 21: EC941 - Game Theory](https://reader034.vdocuments.net/reader034/viewer/2022051219/56816020550346895dcf2121/html5/thumbnails/21.jpg)
The set of rationalizable strategies is {A, B} and {D, E}.Inspection of the best response correspondence shows that there are no pure-strategy Nash Equilibria.
A B C
2, 1
1, 3
1, 1
0, 2
3, 1
1, 1
4, 1
3, 0
2, 2
D E F
1
2
21
![Page 22: EC941 - Game Theory](https://reader034.vdocuments.net/reader034/viewer/2022051219/56816020550346895dcf2121/html5/thumbnails/22.jpg)
With the indifference principle, we find completely mixed strategy N.E. in the game reduced to {A,B}X{D,E}.
Player 1 is indifferent between A and B when2 s2(D) = s2(D) + 3[1- s2(D)] i.e., when s2(D) = 3/4.
A B C
2, 1
1, 3
1, 1
0, 2
3, 1
1, 1
4, 1
3, 0
2, 2
D E F
1
2
22
![Page 23: EC941 - Game Theory](https://reader034.vdocuments.net/reader034/viewer/2022051219/56816020550346895dcf2121/html5/thumbnails/23.jpg)
A B C
2, 1
1, 3
1, 1
0, 2
3, 1
1, 1
4, 1
3, 0
2, 2
D E F
1
2
Player 2 is indifferent between D and E whens1(A)+3[1-s1(A)]=2s1(A)+[1-s1(A)] i.e., when s1(A)=2/3.
The unique Nash Equilibrium of the game has C(s)={A,B}X{D,E} with s1(A)=2/3 and s2(D) = 3/4.
23
![Page 24: EC941 - Game Theory](https://reader034.vdocuments.net/reader034/viewer/2022051219/56816020550346895dcf2121/html5/thumbnails/24.jpg)
A BC
D
0, 3
3, 1
4, 0
1, 0
2, 2
1, 1
0, 2
1, 1
2, 2
E F G H
First, we iteratively delete strictly dominated strategies.
In the first round, only strategy A is deleted.
2, 5 1, 4 0, 2
0, 4
1, 1
0, 1
3, 1
1
24
![Page 25: EC941 - Game Theory](https://reader034.vdocuments.net/reader034/viewer/2022051219/56816020550346895dcf2121/html5/thumbnails/25.jpg)
A BC
D
0, 3
3, 1
4, 0
1, 0
2, 2
1, 1
0, 2
1, 1
2, 2
E F G H
After deleting A, strategy H is strictly dominated.
H is the only strategy deleted in the second round.
2, 5 1, 4 0, 2
0, 4
1, 1
0, 1
3, 1
1
2
25
![Page 26: EC941 - Game Theory](https://reader034.vdocuments.net/reader034/viewer/2022051219/56816020550346895dcf2121/html5/thumbnails/26.jpg)
A BC
D
0, 3
3, 1
4, 0
1, 0
2, 2
1, 1
0, 2
1, 1
2, 2
E F G H
After deleting H, strategy D is strictly dominated.
D is the only strategy deleted in the third round.
2, 5 1, 4 0, 2
0, 4
1, 1
0, 1
3, 1
1
2
3
26
![Page 27: EC941 - Game Theory](https://reader034.vdocuments.net/reader034/viewer/2022051219/56816020550346895dcf2121/html5/thumbnails/27.jpg)
A BC
D
0, 3
3, 1
4, 0
1, 0
2, 2
1, 1
0, 2
1, 1
2, 2
E F G H
After deleting D, strategy E is strictly dominated.E is the only strategy deleted in the third round.The remaining strategies {B,C}, {F,G} cannot be deleted.
2, 5 1, 4 0, 2
0, 4
1, 1
0, 1
3, 1
1
2
3
4
27
![Page 28: EC941 - Game Theory](https://reader034.vdocuments.net/reader034/viewer/2022051219/56816020550346895dcf2121/html5/thumbnails/28.jpg)
A BC
D
0, 3
3, 1
4, 0
1, 0
2, 2
1, 1
0, 2
1, 1
2, 2
E F G H
Inspection of best responses shows that (B, F) and (C, G) are pure-strategy Nash Equilibria.We find mixed strategy N.E. in the game on {B,C}X{F,G}.
2, 5 1, 4 0, 2
0, 4
1, 1
0, 1
3, 1
1
2
3
4
28
![Page 29: EC941 - Game Theory](https://reader034.vdocuments.net/reader034/viewer/2022051219/56816020550346895dcf2121/html5/thumbnails/29.jpg)
A BC
D
0, 3
3, 1
4, 0
1, 0
2, 2
1, 1
0, 2
1, 1
2, 2
E F G H
Player 1 is indifferent between B and C when2s2(F) + [1-s2(F)] = 2[1-s2(F)] + s2(F), i.e. s2(F)=1/2.By symmetry, player 2 is indifferent when s1(A)=1/2.
2, 5 1, 4 0, 2
0, 4
1, 1
0, 1
3, 1
1
2
3
4
29
![Page 30: EC941 - Game Theory](https://reader034.vdocuments.net/reader034/viewer/2022051219/56816020550346895dcf2121/html5/thumbnails/30.jpg)
A Game with Many Players
n people observe a crime.
Each attaches the value v to the police being informed and bears the cost c if calling the police, where v > c > 0.
There are n pure strategy equilibria.
In each one of these equilibria, a player i reports, and the remaining n - 1 players do not.
30
![Page 31: EC941 - Game Theory](https://reader034.vdocuments.net/reader034/viewer/2022051219/56816020550346895dcf2121/html5/thumbnails/31.jpg)
31
Let p be the mixed strategy that a player calls the police.
The symmetric mixed strategy equilibrium p is determined by the indifference condition:
v − c = 0・ Pr{no one else calls}
+ v・ Pr{someone else calls}
v − c = v・ (1- Pr{no one else calls})
v − c = v・ (1 – (1 – p) n-1) or c = – v (1 – p) n-1
p = 1 − (c/v)1/(n−1)
![Page 32: EC941 - Game Theory](https://reader034.vdocuments.net/reader034/viewer/2022051219/56816020550346895dcf2121/html5/thumbnails/32.jpg)
Correlated Equilibrium: Example
DoveHawkDove
Hawk
0, 0
1, -2 -3,-3
-2, 1
In the hawk-dove game, two States confront in a dispute.If a State plays Dove and the other Hawk, the Dove loses the dispute. But if both States are Hawks, a war takes place.
32
![Page 33: EC941 - Game Theory](https://reader034.vdocuments.net/reader034/viewer/2022051219/56816020550346895dcf2121/html5/thumbnails/33.jpg)
DoveHawkDove
Hawk
0, 0
1, -2 -3,-3
-2, 1
There are two pure-strategy Nash Equilibria (H, D) and (D, H), with payoffs (1, -2) and (-2, 1).
There is a mixed strategy Nash equilibrium s(D) = 1/2, which yields payoff of -1 to both players. 33
![Page 34: EC941 - Game Theory](https://reader034.vdocuments.net/reader034/viewer/2022051219/56816020550346895dcf2121/html5/thumbnails/34.jpg)
If the players can play (H, D) and (D, H) with equal probability, their expected payoff is -1/2.
Playing the Nash equilibria (H, D) and (D, H) with equal probability, is a correlated equilibrium.
The players coordinate on a different Nash equilibrium on the basis of a fair coin toss. The coin is the correlating device.
Can the players achieve a higher payoff with a more complex correlating device?
34
![Page 35: EC941 - Game Theory](https://reader034.vdocuments.net/reader034/viewer/2022051219/56816020550346895dcf2121/html5/thumbnails/35.jpg)
Suppose that they resort to a mediator:
1. The mediator randomly chooses between one of the four outcomes (H, H), (H, D), (D, H), and (D, D) using the probability profile p = (pHH, pHD, pDH, pDD).
2. The mediator does not report her choice. She makes a separate, private recommendation h or d, to each one of the players.
For example, if her choice is (H, D), she recommends h to player 1 and d to player 2.
The probability profile p is a correlated equilibrium if all the private recommendations are self-enforcing.
35
![Page 36: EC941 - Game Theory](https://reader034.vdocuments.net/reader034/viewer/2022051219/56816020550346895dcf2121/html5/thumbnails/36.jpg)
For player 1, the private recommendation d is self-enforcing if
U1(D|d) = u1(D,D)p(DD|d) + u1(D,H)p(DH|d) >U1(H|d) = u1(H,D)p(DD|d) + u1(H,H)p(DH|d)
and the private recommendation h is self-enforcing if:
U1(H|h) = u1(H,D)p(HD|h) + u1(H,H)p(HH|h) > U1(D|h) = u1(D,D)p(HD|h) + u1(D,H)p(HH|h)
and the conditions for player 2 are analogous.
36
![Page 37: EC941 - Game Theory](https://reader034.vdocuments.net/reader034/viewer/2022051219/56816020550346895dcf2121/html5/thumbnails/37.jpg)
To find the symmetric correlated equilibrium which maximizes the sum of payoffs, we solve the program:
max [u1(D,D)+u2(D,D)]p(DD)+[u1(D,H)+u2(D,H)]p(DH) +[u1(H,D)+u2(H,D)]p(HD)+[u1(H,H)+u2(H,H)]p(HH)
s.t. Ui(D|d) > Ui(H|d) Ui(D|h) > Ui(H|h) p(DH) = p(HD)
for both players i = 1, 2. 37
![Page 38: EC941 - Game Theory](https://reader034.vdocuments.net/reader034/viewer/2022051219/56816020550346895dcf2121/html5/thumbnails/38.jpg)
38
Substituting in the payoffs, and using symmetry, we obtain:
U1(D|d) = -2p(DH|d) > U1(H|d) = p(DD|d) - 3p(DH|d)
U1(H|h) = p(HD|h) - 3p(HH|h) > U1(D|h) = -2p(HH|h).
Expanding the conditional probabilities, and simplifying
- 2p(DH) > p(DD) - 3p(DH)
p(HD) - 3p(HH) > -2p(HH).
Note that this is a set of linear inequalities.
![Page 39: EC941 - Game Theory](https://reader034.vdocuments.net/reader034/viewer/2022051219/56816020550346895dcf2121/html5/thumbnails/39.jpg)
We solve for the optimal correlated equilibrium:
max 0p(DD)+2(1-2)p(DH) - 6p(HH)s.t. p(DH) = p(HD)- 2p(DH) > p(DD) -3p(DH) p(HD) -3p(HH) > -2p(HH)
Simplifying, we obtain: p(DH) = p(HD) > p(DD) and p(HD) = p(DH) > p(HH).
The solution is p(HH)=0, p(HD)=p(DH)=p(DD)=1/3.
The expected payoff of both players i = 1, 2 is:
Ui(p) = 0p(DD) + p(HD) - 2p(DH) = -1/3.
39
![Page 40: EC941 - Game Theory](https://reader034.vdocuments.net/reader034/viewer/2022051219/56816020550346895dcf2121/html5/thumbnails/40.jpg)
Correlated EquilibriumDefinition A correlated equilibrium in a game G=(I, S, u) is a probability distribution p over the set S such that, for any player i, and all her strategies si
Ss-i [ui (si, s-i) p(si, s-i)] > Ss-i [ui (s’i, s-i) p(s’i, s-i)], for all s’i.
Proposition Any correlated equilibrium is also a mixed strategy Nash Equilibrium.
40
![Page 41: EC941 - Game Theory](https://reader034.vdocuments.net/reader034/viewer/2022051219/56816020550346895dcf2121/html5/thumbnails/41.jpg)
Summary of the Lecture
Mixed Strategies
Nash Equilibrium and Rationalizability
Correlated Equilibrium
41
![Page 42: EC941 - Game Theory](https://reader034.vdocuments.net/reader034/viewer/2022051219/56816020550346895dcf2121/html5/thumbnails/42.jpg)
Preview Next Lecture
Bayesian Games and Nash Equilibrium
Bayesian Games and Information
Cournot Duopoly with Private Information
Public Good Provision with Private Information
42