Professional Documents
Culture Documents
Problem Set 10 Solution, 2021
Problem Set 10 Solution, 2021
In this set, we are trying to go over the versions of the Prisoner’s dilemma we went over in
lectures.
Cooperate Defect
What we are trying to prove is that as we increase the payoffs of both players from mutual
cooperation (underlined above), more people will cooperate since the level of altruism above
which people decide to cooperate goes down.
a. What are the Nash equilibria of the game above when played by two Egoistic
players? Explain your answers.
The game above has a unique Nash equilibrium in which both players
defect. To show that it is a Nash equilibrium we have to go over each of
the players and assume that they know that the other defects. Under this
assumption we check that it is a best response for them to defect. If it is
true for all players we have a Nash equilibrium. If not true, and one or
more of the players get a strictly higher payoff from cooperating, then
this is not an equilibrium.
Focusing on the row player, we consider whether she is best responding
to the column player defecting: she gets 2 from defecting and only one
from cooperating so she is best responding.
Focusing on the column player, we consider whether she is best
responding to the row player defecting: she gets 2 from defecting and
only one from cooperating so she is best responding.
To see that it is the only equilibrium there are two ways to answer:
The quick way to answer: In the game above, for each player, the action
defect strictly dominates the action cooperate. No rational player ever
plays a strictly dominated action. In other words, cooperate is never a
best response (for whatever belief the player might have about what her
opponent will do). Therefore, there cannot be any Nash equilibrium in
which a player plays cooperate!
b. Write down the matrix of the game above when two Altruistic players play this game.
What are the pure strategy Nash equilibria of this game? Explain your answers.
Now each player puts a weight of 1/7 on the payoff of her opponent. The
new game table will be:
Cooperate Defect
Cooperate Defect
c. Write down the matrix of the game above when two Fully Altruistic players play this
game. What are the pure strategy Nash equilibria of this game? Explain your
answers.
Now each player puts a weight of 1 on the payoff of her opponent. The
new game table will be:
Cooperate Defect
Cooperate Defect
Note that the payoffs have changed and now defect does not strictly
dominate cooperate for each player.
Cooperate Defect
p 1-p
Cooperate
6,6 8,8
q
Defect
8,8 4,4
1-q
In equilibrium, because they mix between the two actions, they have to
get the same expected utility from both actions:
q×6+(1-q)×8=q×8+(1-q)×4
Assume now that we change the game by only increasing the payoffs from mutual
cooperation. The new game is:
Cooperate Defect
d. What are the pure strategy Nash equilibria of the game above when played by two
Egoistic players? By two Altruists? By two Fully Altruistic players? Explain your answers.
Egoistic players:
The game above has a unique Nash equilibrium in which both players
defect. To show that it is a Nash equilibrium we have to go over each of
the players and assume that they know that the other defects. Under this
assumption we check that it is a best response for them to defect. If it is
true for all players we have a Nash equilibrium. If not true, andone or
more of the playersget a strictly strictly higher payoff from cooperating,
then this is not an equilibrium.
Focusing on the row player, we consider whether she is best responding
to the column player defecting, she gets 2 from defecting and only one
from cooperating so he is best responding.
Focusing on the column player, we consider whether she is best
responding to the row player defecting, she gets 2 from defecting and
only one from cooperating so she is best responding.
To see that it is the only equilibrium there are two ways to answer:
The quick way to answer: In the game above, for each player, the action
defect strictly dominates the action cooperate. No rational player ever
plays a strictly dominated action. In other words, cooperation is never a
best response (for whatever belief the player might have about what her
opponent will do). Therefore, there cannot be any Nash equilibrium in
which a player cooperates!
Altruistic players:
Now each player puts a weight of 1/7 on the payoff of her opponent. The
new game table will be:
Cooperate Defect
6.3+6.3/7,
Cooperate 1+7/7,7+1/7
6.3+6.3/7
Cooperate Defect
(6.3×7+6.3)/7,
Cooperate 2,50/7
(6.3×7+6.3)/7
Note that the payoffs have changed so that defecting does not strictly
dominate cooperating for each player. This is because
(6.3×7+6.3)/7 > 50/7!
Full Altruists:
Now each player puts a weight of 1 on the payoff of her opponent. The
new game table will be:
Cooperate Defect
Cooperate Defect
Note that the payoffs have changed and now cooperate strictly
dominates defect for each player. Therefore the only Nash equilibrium is
for both players to choose cooperate.
e. In light of the results above, how can you explain the experimental results we saw in
lectures in which as the payoff from mutual cooperation was increased, the fraction of
subjects in the lab that cooperated increased?
To explain this in view of the above answers, we can conjecture that the
population of lab subjects are of different altruism levels, so that some
are egoistic, some are altruistic and some are fully altruistic. When the
payoff from mutual cooperation was 3, we saw that only the fully
altruistic were cooperating.
Therefore, our results above are consistent with the observations in the
lab experiment.