1-1 What is Game Theory?

question number one is about the political campaigns between the Democrats and the Republicans.
What determines the policies of the two parties?
Republicans and the Democrats. Well, this is a very general and vague question.
Think about what, how you can approach this question. It's not so obvious where you start.
But game theory provides you a unified way of solving all social problems including this one.
So, you formalize political struggle between the two parties by means of a simple mathematical
model. And then you apply the basic solution concept of game theory and then you get an answer.

Question number two is about the traffic on the road, There

are two cities here and there. Say city X and city Y. And those are the existing roads. And the
question is what happens if we construct a new road coming from X to Y? How does it affect the
traffic on the existing road? And how you can save travelling time from X to Y? This is a very
important policy question. Again, this is a strategic situation. If all the drivers jump to the new
bypass then this bypass is too crowded. And maybe you would like to use the existing road. So what
is best for you crucially depends on what other drivers do. This is a perfect example of game
theoretic or strategic situation. Again, you formulate this social problem as a mathematical
model game and then you apply the basic solution concept. And then you get the answer.

So the struggle for existence of animals and plants, it's a kind of game. What is best for one specie may
depend on what other specie behave. So you can apply the concept of game theory to biology and
the branch of biology which uses game theory is called evolutionary game theory.
What about the relationship between game theory and computer science? Computer science used to
analyze the calculation performed in a single computer. But now computers all over the world are
interconnected with with network. So those computers are collaborating or often competing to perform
their calculations. This is also a strategic situation, and game theory can be fruitfully applied to
computer science.
Well, I am an economist, and I do research in economic theory. But I have been collaborating with
sociologist and computer scientist, and I have been interacting with biologist.
Game theory provides a common language which facilitates interdisciplinary research in all those
different fields.
1-2 Modelling Social Problems as a "Game"

You need to simplify the reality. And you have to capture the essence of reality in the simple
model of game

after formulating a social situation as a mathematical model of game, then you have to find a
solution to the game, you have to predict how people behave. So then, game theory takes a unified
approach. All those social problems have the following two features.
Number one, individuals try to do their best against others.
Number two, but they try to do their best under a certain set of rules.

von Neumann, a computer scientist and mathematician and theoretical physicist,

and Oskar Morgenstern a professor of economics at Princeton University.
The Theory of Games and Economic Behavior, 1944.
That was the beginning of game theory.
1-3 In Search for the Governing Principle
1-4 Concerns About a Mathematical Theory of Human Behavior

Three skeptic's points of view

- free will defeats any attempt to predict human behavior by a mathematical formula
-we don't need a mathematical theory to explain or predict human behaviour because eventually, you can ask why
do you do that
-I have never seen or heard that game theory works

Around 1939
Then, game theory was created, and now we have some evidence that mathematical approach actually works
very well.

Why do we need the game theory, rather than the theory of maximization, to analyze a strategic situation?

Because many players are not familiar with the theory of maximization.

Because we need an additional theory to analyze how one player anticipates the other players' behavior.

Because human emotions cannot be captured by mathematics.

Because the theory of maximization is too difficult to handle.

Correct

1-5 Let's Play a Game

Who has an advantage? What is the winning rate or winning probability of each player?
Do you have any prediction about the distribution of cards you choose?
John Nash Discovered the Governing Principle

The question is if we can find a unified general theory to find the solution to all of those social
problems. Is there any single solution concept which can be applied to all those social interactions?

suppose the surface of coffee represents the set of all possible human behavior in a social problem.
This was really an ingenious idea of John Nash.

So by using advanced mathematics called topology, Nash discovered or found that any social
problem has a Nash equilibrium point. Nash discovered that every social problem has a stable point
where all individuals are doing their best against others. And this is called Nash equilibrium.

So first, you formulate a social situation as a game, a mathematical model of a game, by specifying
players, strategies, and payoffs. And then you use your mathematics to find out something like a
vortex point, where all individuals are doing their best against others. And then you get a prediction
in every single social problem. That's called Nash equilibrum.
1-7 Nash Equilibrium

A Nash equilibrium is the combination of strategies in which no single player can obtain a higher
payoff by changing his/her strategy unilaterally. They are taking mutually best response.

Practice Quiz 1.2

TOTAL POINTS 2
1.Question 1
Which of these questions was answered by John Nash?
How can ad hoc research be conducted effectively?
How can the vortex of coffee cup be found?
How can the unified solution concept, which can be applied to analyze a variety of social problems, be
formulated? Correct
How can people’s expectations about future events be analyzed?

2.Question 2
What is the definition of the Nash equilibrium?
A combination of strategies that are best replies to each other. Correct
A combination of strategies that is most desirable for players.
A combination of strategies where players cannot increase their payoffs by deviating together.
A combination of strategies where each player is minimizing other players’ payoffs.
Traffic Game in Reality
Location Game (Hotelling’s model)
1-10 Policies of Two Parties

median over distribution is determined by the following way.

So, half of the population is here and the other half of the population is here.

What was the accuracy of game theoretic predictions of the traffic allocation on various roads around
Hamamatsu city?

Nash equilibrium almost perfectly predicted actual traffic.

Nash equilibrium predicted actual traffic reasonably well. Correct
Nash equilibrium predicted actual traffic poorly.
Graded Quiz 1 LATEST SUBMISSION GRADE 100%
TOTAL POINTS 7

1.Question 1
Which of the following best describes a “strategic situation”?
You cannot trust your opponent.
Finding the optimal strategy needs non-trivial calculation.
What is best for you depends on what others do.
You and your opponent are in serious conflict of interest.

2.Question 2
A strategic situation can be formulated as a mathematical model of a “game” that specifies three sets of
items. Choose the three sets of items from the following.
(A) The optimal strategy
(B) Players
(C) Referee
(D) Strategies
(E) Social value
(F) Payoffs
(G) Penalties
B, D and F

3.Question 3
Rational players’ behavior in a game cannot be analyzed by a simple application of mathematics (using
the theory of maximization). What is the main reason for this?
The maximization problem often has no solution in a game.
Mathematical theory of maximization fails to capture players’ skills.
Mathematical theory of maximization fails to capture players’ psychology.
If we just assume that each player maximizes his/her payoff, we fail to pin down how each one
anticipates the other players’ behaviors.

4.Question 4
Which of the following actions best describes that you are rational?
You choose the best available option.
You can suppress your emotions.
You can predict the future.
You exploit others whenever possible.
1 point
5.Question 5
What are the general properties of Nash equilibrium? Choose all applicable properties.
The sum of players’ payoffs is maximized.
Each player assumes that other players take the worst strategies for their opponents.
Players are making mutual best replies.
No players can increase their payoffs by deviating by themselves.

6.Question 6
Consider the traffic game explained in Lecture 7 in Week 1. Suppose the shortest route (whose length is
200) is a new bypass.

If we construct this new bypass, the traveling time* from City X to City Y is reduced.
* Traveling time = (Route length) + (Number of cars on the route)
By comparing the Nash equilibria before and after the construction of the new bypass, compute the
reduction of the traveling time achieved by the new bypass. Enter your answer in numerical form
(unsigned number) in the box below.
e.g. If the result of your computation is 500, then just enter 500 in the box. If -500 comes out, enter 500
after removing the negative sign "-".
75
(375-300)

7.Question 7
What is the significance of John Nash’s discovery?
He discovered the infinite regress problem.
He found a unified solution concept that can be applied to a wide range of games.
He discovered that the surface in a coffee cup has a vortex.
He showed how to do ad hoc analyses of strategic situations.
Question 1
[Answer] There is no Nash equilibrium in which each vendor chooses the location not
probabilistically but deterministically.

[Explanation] Nash equilibria in which each vendor does not choose the location probabilistically are
called pure strategy Nash equilibria. The precise definition is given in Module 2. When there are
three competing players in Hotelling’s location game, there is no pure strategy Nash equilibrium.

Question 2
[Answer] [1/4, ¼,3/4,3/4] constitutes a Nash equilibrium.

[Explanation] When there are four competing players in Hotelling's location game, only this
combination of locations constitutes a pure strategy Nash equilibrium.

Question 3
[Answer][1/2,1/2] constitutes a Nash equilibrium.

[Explanation] In the game presented in the lecture, each party’s objective is to win the election (i.e.,
parties are office-motivated). The median voter theorem in the model with office-motivated parties
was proved by Downs (1957). The game we solved in the problem is different from the Downs’
model in that parties are policy-motivated. This type of setting was explored by Wittman (1973).
Interestingly, the median voter theorem still holds.
2-1 Nash Equilibrium and the Prisoner’s Dilemma

- a mutual best reply

- no single player can increase his payoff by deviating by himself

Games without any timing are sometimes called simultaneous move game - players take their
actions simultaneously, and the game ends immediately
= static game

Why is it called prisoner's dilemma? Well, so if each individual seeks his own interest,
rational behavior of each individual leads to Nash equilibrium, mutual defection.
On the other hand, mutual cooperation is better for the group of two players as a whole.
Group rationality indicates that mutual cooperation is even better.
So prisoner's dilemma game clearly shows that group rationality- what is best for the group what is
best for the society, Is not always equal to what is best for each individual.
This gap or difference between group rationality and the individual rationality is very important.

ex ante /ɛks ˈanti/ adjective · adverb

based on forecasts rather than actual results.
"an ex ante estimate"
2-2 Coordination Game and Self-Fulfilling Prophecy

Those two formats are incompatible. The society has at least two equilibria.
In one equilibrium, everybody uses HD DVD. So here, HD DVD became de facto standard of the
new generation DVD.
The other equilibrium everybody uses Blu-ray, so there was a struggle for de facto standard, okay.
There is no guarantee that the best format always becomes the de facto standard!

a perfect example of the phenomenon called self-fulfilling prophecy. So you can imagine that there is
a prophet saying that behold Blu-ray users, the end is near, repent and switch to HD DVD, and if
everybody believes what he says, everybody buys HD DVD and HD DVD is becoming more
and more desirable, and eventually people follow the prediction of this person. This is what is called
self-fulfilling prophecy, but since the society has another equilibrium, Blu-ray equilibrium, there is
another self-fulfilling prophecy, okay? So, if the society has multiple Nash equilibria,
we have multiple self-fulfilling prophecies. So, this prophet here is saying that, hey, no- pay no
attention to this stupid guy. Stick to Blu-ray, buy Blu-Ray, and let's see what actually happened,
okay, eventually, people believed this guy and Blu-ray dominated the market, okay.

So let me sum up. Society may have multiple Nash equilibria in this case two equilibria.
Everybody uses HD DVD, everybody uses Blu-ray, and eventually Blu-ray prevailed,
okay, and which one eventually prevails depends on people's expectations.
So ex ante, it's really hard to predict which one eventually dominates, and also,
there is no guarantee that the best format always becomes the de facto standard.
2-3 Market Competition

if there are lots of firms, and if price is above ten cents, it's profitable to produce a pencil,
so lots of firms would like to produce pencils.

So, if price
p> ten cents, supply is infinite
p< ten cents, it doesn't make sense to produce pencil, so supply is zero,
p= ten cents, then you are indifferent between, you don't mind producing pencil.

So therefore, economics textbook says that supply is a flat curve.

Fhe economics textbook says that price and quantity in perfectly competitive market where there are
lots of firms, perfect competitive equilibrium is given by the intersection between demand and
supply. So price and quantity is given by this intersection point, okay.
This is what is called the perfectly competitive outcome or market equilibrium.

So economics textbooks says that large number of firms implies perfect competition.
So let's examine by using game theory if this is actually true.

So, let's calculate the perfect competitive equilibrium intersection between demand and supply.
The quantity of perfectly competitive market partly depends on the unit cost of production of pencil.
So Nash Equilibrium says the following things. If there are few firms in the market, price is very high
and quantity is small, but as the number of firms increase, price goes down and eventually it
converges to competitive, perfect competitive market equilibrium, okay?
So, large number of firms actually implies perfect competition. This is the prediction by Nash
equilibrium.
So by using the concept of Nash equilibrium, and by using the power of math to find Nash
equilibrium, we have derived the Law of Market Competition in economics.
What are the properties of the Prisoner’s Dilemma game? Choose all applicable properties.
(A) Mutual cooperation is better than mutual defection for all players.
(B) Mutual cooperation and mutual defection are both Nash equilibria.
(C) For any given action of your opponent, defection is always better than cooperation.
(D) Nash equilibrium does not exist.
Answer: A and C
Why Do People Come to Play Nash Equilibrium? Part I
2-5 Why Do People Come to Play Nash Equilibrium? Part II

if players don't know how to play a game, a reasonable thing for them to do is to get together and
talk about what they should do. Ex ante, when they don't talk, it's not so clear which equilibrium to
play. So naturally in this situation they just talk.

obviously, all those drivers didn't have any opportunity to get together and talk about which
equilibrium to play.

So one day I took this route and it was too crowded, and I switched to better route.
And by means of this trial-and-error adjustment I got better expectations, I got better beliefs about
what is the traffic of each route? And eventually I'm choosing optimal route.
And then people play Nash equilibrium of this traffic game.
2-6 Why Do People Come to Play Nash Equilibrium? Part III

Experimental Evidence

Players 2 is a winner

obviously this is not a Nash Equilibrium

this is a Nash Equillibrium (payoff is 5)

again it is a Nash Equillibrium (more generally)

So, accumulating experience, they learned how to play Nash equilibrium here. (8th round)
2-7 Stylized Facts and Nash Equilibrium
People might not play Nash equilibrium, and I'm going to tell you
if game theory is useful in those situations.

if the situation is not Nash equilibrium, there's at least

one player who can deviate to increase his payoff. So, the situation is not stable, and adjustment
process doesn't stop there, and goes on.
So if the adjustment ever settles on a certain point, that should be Nash Equilibrium, okay,
that's fine, but there is no guarantee that, you know, this adjustment process always converges,
so in that case, you know, the behavior might be chaotic, and even if adjustment
process eventually does converge to a Nash equilibrium, it may take a very long time.

this is a Nash equilibrium. If all other people are following this convention, it's best for you to follow
the same convention. So if you are standing on the right-hand side, you will be harassed by
customers in a hurry, so this is a Nash equilibrium. If you deviate, you will lose.

if you find a stable mode of behavior, which is repeatedly observed in a society, a stylized fact that
is highly likely to be a Nash equilibrium. Very important goal of any social science is to explain
stylized fact and the stylized facts are likely to be Nash equilibria. So this is probably one of the most
important reasons why Nash equilibrium is useful in social sciences.
Practice Quiz 2.2
TOTAL POINTS 2
1.Question 1
Choose the correct combination of words to fill the blanks in the following sentences explaining the
conditions for Nash equilibrium to be established. Note that blanks with the same alphabet are filled
with the same words.

-----

Rationality alone (A) guarantee that Nash equilibria are played. To play Nash equilibria, players
need (B). One way that makes (B) is (C). If players can reach (D) in (C), they can play a Nash
equilibrium.

-----

A= does , B = unlimited capacity of reasoning , C = trial and error adjustment, and D= self-
fulfilling prophecy.
A= does not , B = unlimited capacity of reasoning, C = pre-play communication, and D= self-
enforcing agreement.
A= does not , B = correct beliefs, C = pre-play communication, and D = self-enforcing
agreement. Correct

A= does , B = correct beliefs, C = trial and error adjustment, and D= self- fulfilling prophecy.

2.Question 2
In general, rationality alone is insufficient for players to choose Nash equilibrium strategies. What
additional condition assures that they play a Nash equilibrium?

Players know the payoffs of all players.

Each player has correct beliefs about the strategies of other players. Correct
Trial and error adjustment.
Unlimited capacity of calculation.
2-8 Make Yourself Unpredictable: Mixed Strategy Equilibrium

I will explain that sometimes, you know, taking random behavior,to make yourself unpredictable is
very important in some social situations. And it is about what is called mixed strategies and mixed
strategy equilibrium.

no mutual best reply point. But in this game, it's very important to make
yourself unpredictable. If you always choose stone, you lose. So in this game of rock, paper,
and scissors it's very important to be unpredictable. And game theory can formulate the situation
where playing random behavior is the best. So you should randomly choose your action.

Nash equillibrium in random strategies

2-9 Sports Games and Game Theory

researchers also discovered that game theoretic prediction based on mixed

strategy works surprisingly well in other professional games such as tennis games.
Nash Equilibrium Exists in All Games

The first condition, there are finitely many players, this condition is always, almost always satisfied.

finite slots
 a mixed strategy Nash equilibrium.

The existence of Nash equilibrium, and the existence of vortex point, follows from the same
mathematical fact, which is known as Fixed Point Theorem.
TOTAL POINTS 7
1.Question 1
Consider the following two-player game with three actions for each player. How many Nash equilibria are there?
(If you are familiar with the concept of mixed strategies, do not count the mixed equilibria.)

1 2 3 4
c,x b,y a,z

2.Question 2
In the coordination game without correct beliefs, which of these can occur?

Players actually play a Nash equilibrium, and which equilibrium is played is unpredictable.

Belief disagreement may leads to non-Nash outcomes, and its realization is unpredictable.

Players actually play a Nash equilibrium, and which equilibrium is played is predictable.

Belief disagreement may leads to non-Nash outcomes, and its realization is predictable.
3.Question 3
Consider a game represented by the following table.

Which of the following statements is true?

In a Nash equilibrium, both players choose C.

In a Nash equilibrium, both players choose B.

In a Nash equilibrium, both players choose A.

There is no Nash equilibrium.

4.Question 4
Concerning the above question 3, which of the following statements is correct?

This game is similar to the prisoner’s dilemma in that Nash equilibrium is a better outcome than every other
outcome for all players.

This game is similar to the prisoner’s dilemma in that it has a better outcome than Nash equilibrium for all
players.

This game is similar to the coordination game in that there are two Nash equilibria.

This game is similar to the coordination game in that each player’s objective is to match his/her behavior with
the others.

5.Question 5
What kind of insights does the Coordination Game provide? Choose all applicable statements.

There can be many Nash equilibria in a game.

The popular computer keyboard design was inherited from typewriters.

Players can always coordinate on the optimal outcome.

The de facto standard of a new technology may not be efficient.

A game might have good and bad Nash equilibria (the former are better than the latter for everyone).

There is a unique self-fulfilling prophecy.

6.Question 6
Which game has no mixed Nash equilibrium, i.e., no Nash equilibrium where at least one player chooses random
behavior? Choose the correct game with a correct reason.

The coordination game, because payoffs for each player are the same across two Nash equilibria.

The matching pennies, because the result of a coin-toss trial can be perfectly predicted by today’s science.

The prisoner’s dilemma, because defection is best for both players and they do not consider that mutual
cooperation can be attained.

The battle of the Sexes, because players have made a promise where to visit during the next holiday.
7.Question 7
What is the property of a self-enforcing agreement?

Agreement is fulfilled by a moral judgment.

Agreement is fulfilled by a legal contract.

Agreement is fulfilled by a law enforcement agent.

Agreement is fulfilled without imposing penalty or reward.

5. Which sentence in the following explains the trial and error adjustment?

Played it many times so that better understanding is achieved.

There is Nash equilibria for k=0, and for k=3

1/4 X=Y Y=Z p, q, 1-p-q

3-1 Digression: The Card Game Revisited

Who has as advantage, red or black?

What is the winning rate of each player? What's the probability of winning of red and black?
can you say anything about the distribution of cards of each player?
3-2 Digression: How You Played the Card Game and Addressing the
Concerns about Game Theory

Poll Results
What is your expectation on the accuracy of game-theoretic prediction of the winning rate, taking the red
player's winning percentage?
-So-so (Red player's winning rate between 0.3 and 0.5) 14%
-Reasonably good (Red player's winning rate between 0.35 and 0.45) 61%
-Amazingly good (Red player's winning rate between 0.39 and 0.41) 25% 1,160 responses

1= ace (so bigger rate)

1.free will <-> people are attracted to the behavior that is best for them, i.e.,
given that other players are choosing equilibrium strategy, it's best for you to follow this equilibrium

2. why did you do that <->

game theory can uncover the mechanism operating behind your instinctive behavior

3. I’ve never heard it works (name me one proposition in all of the social sciences which is both true
and non-trivial – Stan Ulam/Paul samuelson) <->
This card game is one of those examples.
3-3 “Payoffs” in a Game: What Exactly Are Those Numbers?
case 1

case 2

No right or wrong answer

Practice Quiz 3.1
TOTAL POINTS 1
1.Question 1
Suppose your opponent is choosing actions A and B randomly, with equal probability. Your choice is
either X or Y, and you get a prize according to the following table.

Hence, if you choose X, you get $10 or $0 randomly, and on average you can
earn 12$10+12$0=$5. In contrast, choice Y gives you $5 for sure.
Suppose you hate risk and therefore prefer $5 for sure. Your behavior can be captured if we assume
that (i) your payoff is your “utility” of the prize and (ii) you maximize the expected value of
your utility. Let your payoffs (utilities) be given by the following table.

What is the condition that those numbers u(10u(0), and u(5) should satisfy?

1 / 1 point
1/2 u(10) + 1/2 u(0) = u(5)
1/2 u(10) + 1/2 u(0) < u(5) Correct
1/2 u(10) + 1/2 u(0)> u(5)
1/2 u(10) - 1/2 u(0)= u(5)
3-4 What Does it Mean That a Player is Rational?

clear-cut /klɪəˈkʌt/ adjective.

1.sharply defined; easy to perceive or understand.
"we now had a clear-cut objective"
Similar:definite distinct clear well defined sharply defined precise specific explicit unambiguous black and white
hard and fast unequivocal straightforward striking obvious marked decided sharp crisp stark cut and dried
Opposite: blurred indistinct vague
2.(of an area) from which every tree has been cut down and removed.
"clear-cut forests are unappealing for recreation"
verb cut down and remove every tree from (an area).
"colonizers clear-cut large jungle tracts"

3-5 Domination: Strategies That Are “Obviously Good or Bad”

 Domination Definition:
 Strategy A (strictly) dominates strategy B when strategy A yields higher payoffs than strategy B
regardless of what the opponent does.
 Strategy A weakly dominates strategy B when strategy A never yields strictly lower payoffs and
sometimes yields strictly higher payoffs than B.

end-slots
3-6 Common Knowledge of Rationality
Question 1
Which of these is a property of the Battle of the Sexes?
For any given action of your opponent, the alternate action is the best choice.
There are two non-random Nash equilibria, and one is better for Man (player 1) and the other is
better for Woman (player 2). Correct
There are two non-random Nash equilibria, and one is better than the other for all players.
Nash equilibrium does not exist.

2.Question 2
What does it mean to say that your strategy X strictly dominates your strategy Y?
Strategy X is better than strategy Y, if other players are making their best replies.
Y gives the worst possible outcome for you.
Strategy X always gives a strictly higher payoff to you than Y does, irrespective of other players’
strategies. Correct
Choosing X is always optimal for you.
3-7 Low Rationality: What Happens if Players Are Not Very Smart?
Eventually, A finds that the end slots are bad
3-8 Game Theory Under Zero-Intelligence: Biological Evolution

at the end

Survival of the fittest, the outcome of evolution

mutation, D- new strategy

A survives

3-9 Fig Wasps Play a Nash Equilibrium

When fraction of flying male is small, then flying males have better chance of getting lots of
offsprings.
So vertical axis measures the fitness of each strategy, flying and staying, and it’s the number of
offsprings.
When the fraction of flying males is small, then flying males have a better chance of getting many
offsprings.
Where flying males are rare, a flying male have a higher fitness than staying males.
And therefore, the fraction of flying male is going to increase.

This arrow indicates the nature of natural selection When the fraction of flying male is small.

The situation is opposite when fraction of flying male is large flying males have less offspring.
actual data

we can conclude that the fig wasps playing the Nash equilibrium of a game played by genes.
This is an example of an application of game theory to biology.
Practice Quiz 3.3
Why is game theory useful to analyze animal behavior, which has zero intelligence to reason?
Because...

Because the outcome of biological evolution can be regarded as a Nash equilibrium of a game
played by genes. Correct
Because every creature has an instinct to destroy other species, and this is a strategic situation.
Because each gene can change its strategy flexibly, and thus, their interactions lead to a Nash
equilibrium.
Because animals actually have low-intelligence, and thus, game theory can be applied partially.

2.Question 2
Animals and insects have very limited intellectual capacity for any rational reasoning. Yet, their
behavior can sometimes be explained by game theory. What is the main reason for this?

Although each individual has limited intellectual capacity, a group of animals or insects can
sometimes conduct rational calculations.
When behavior is genetically determined and inherited, the share of a successful behavior
(“successful” in the sense that it produces many offspring) increases over time. Correct
The concept of Nash equilibrium takes into account the intellectual capacity of players.
Mutation introduces the intellectual capacity for rational reasoning in the long run.

Correct: Player 2 does not play R

3. In the game in the above question 2, suppose Player 2 is rational and she/he knows that Player 1 is also rational.
Which of the following is true?

Since Player 1 may not know that Player 2 is rational, Player 1 might choose A. Hence, Player 2 might want to
choose X, which is optimal when Player 1 chooses A.

Player 2 knows that Player 1 never chooses A. Given that, Z is always best for Player 2. Therefore, Player 2
chooses Z.

Since both players are rational, they choose Nash equilibrium (C, Z).

Player 2 knows that Player 1 knows that Player 2 is rational.

4. Under which of the following conditions, can we conclude that players choose the Nash equilibrium (C, Z) of the
game in the question 2?

Both players are rational. Both players know that they are both rational. Player 2 knows that both players
know that they are both rational.

Both players are rational. Both players know that they are both rational.

Both players are rational.

Both players are rational. Both players know that they are both rational. Both players know that they know
that they are both rational.
Choose the correct combination of words to fill the blanks in the following explanation about the evolutionary
game theory.
-----
5. In evolutionary game theory, strategies are genetically programmed characteristics that are determined by
genes. Players are (A). Payoffs represent (B). These represent the basic concepts in evolution, or (C).
-----

A= certain groups of animals, B = number of offsprings, and C= natural selection

A = also genes, B = result of hunting, and C = the divine will

A = also genes, B = number of offsprings, and C= natural selection

A = certain groups of animals, B = result of hunting, and C= Darwin’s law

Which of the following is the correct statement regarding payoffs in game theory?
1 point

When there is uncertainty, payoffs describe monetary profits obtained by each action.

When there is uncertainty, payoffs represent utility from each outcome, which describes players' attitude to
risks.

When there is no uncertainty, payoffs describe monetary profits obtained by each action.

When there is no uncertainty, payoffs uniquely describe values obtained by each action.

Module 3 Optional Challenge Problems

TOTAL POINTS 4
1.Question 1
Consider the following experiment.
There is a pot containing 60 balls. Twenty balls out of these sixty are red, and the remaining forty balls are
either blue or white. In this experiment, a student picks up one ball from the pot under the following
conditions:
The student knows that the total number of balls is 60 and that the number of red balls is 20.
The student does not know the number of blue balls or white balls.
A student (subjectively) believes that the number of blue balls is b.
Note that the student is the expected utility maximizer, and that the student's expected utility function
u is u(∙)=(∙).
Under this condition, the student was asked the following question:
[Question A to the student] Which lottery do you prefer?
A-1: Receive $K irrespective of which ball you pick up.
A-2: Receive $100 if you pick up either a red or blue ball, or zero reward otherwise.
When b=10, the student answered that “I’m indifferent between A-1 and A-2, because the expected
utilities of A-1 and A-2 are the same.”
What was the value of K in this situation? Enter the value of K in the answer column below.
25

2.Question 2
(Continued from Question 1, but b is unknown in this question)
Next, the student was asked the following two questions:
[Question B to the student] Which lottery do you prefer?
B-1: Receive $100 if you pick up a red ball, or zero reward otherwise.
B-2: Receive $100 if you pick up a blue ball, or zero reward otherwise.

[Question C to the student] Which lottery do you prefer?

C-1: Receive $100 if you pick up either a red or a white ball, or zero reward otherwise.
C-2: Receive $100 if you pick up either a blue or a white ball, or zero reward otherwise.

The student answered that s/he strictly preferred B-1 to B-2 in Question B, and strictly preferred C-2 to
C-1 in Question C.
Based on the above answers of the student, assuming that the student maximizes her/his expected
utility, let us calculate what the student believes to be the value of b.
First, we can find the range of b from the answer to Question B:
The student prefers B-1 to B-2 if and only if b < (i).
Second, we can also find the range of \mathbf{b}b from the answer to Question C:
The student prefers C-2 to C-1 if and only if b > (ii).
What are (i), and (ii) in the sentences above? Enter your answers to (i), and (ii) in this order,
separated with a space, in the answer column below.
2 5
20 20

3.Question 3
Consider the game explained in Lecture 2-3, where there are N firms and all firms choose the amount of
production simultaneously. The rest of the conditions are the same as explained in Lecture 2-3. Note
that the firms do not choose negative amounts of production.

Now let’s think about the case where 2N=2 (there are two firms: A and B), a=5, b=1 and c=2. The firm
A decides its amount of productionbqA according to the following conditions:
The firm A is rational.
The firm A knows that the firm B is rational.
From the options below, identify qA that the firm A never chooses. Select all that apply.
 0 ¼ 1/2 3/4 1 5/4 3/2 7/4 2
the remaining are the best responses around NE
4.Question 4
Suppose that the firms A and B are rational, and the rationality of the firms is common knowledge.
Calculate the amount of production of the firm A: qA∗ in this case, and enter the value of qA∗ in the
answer column below. Note that qA∗ can be any non-negative real number.
1

Question 2
[Answer] 20 20

[Explanation] As you can see, the two answers contradict each other. This means that the expected
utility theory does not explain this student’s choices. Furthermore, many experiments find that
subjects often choose B-1 over B-2, and C-2 over C-1. This paradox is well known in economics and
is called the Ellsberg paradox, found by Ellsberg (1961). To explain this anomaly, alternative
theories (e.g., Maxmin expected utility theory) have been developed.

Question 3 and 4
A step-by-step solution for Question 3 and 4 is presented below.

Iterated Elimination of Strictly Dominated Strategies IESDS

[Step 0] Examine firm A's profit function

We start with observing how the firm A's profit function looks like, as shown in the figure below.
Given qB, the firm A's profit function is a quadratic function which takes the highest value when qA is
the best response to qB. Since this is a quadratic function, the profit monotonically increases as qA
becomes closer to the best response qA∗(qB). This property plays a key role in the following
analysis, so keep this in mind.

[Step 1] Consider the first condition: "The firm A is rational."

Suppose that the firm A does not know that the firm B is rational. This means that the firm B might
produce any amount. Given this, determine the amount of production the firm A never chooses, so
long as A itself is rational. As discussed in the lecture, the firm A never chooses strictly dominated
strategies. Recall that strategy x is strictly dominated by y, if y provides a higher payoff
than x for any strategy of the opponent.

The figure below demonstrates that qA>1.5 is strictly dominated by qA=1.5 (so that the firm A never
chooses qA>1.5 if it is rational).
The figure above shows the best response function of the firm A. Let’s compare the firm A's two
strategies, qA=1.5 vs. qA>1.5. For any action taken by the firm B, the best response of the firm A is
smaller than or equal to 1.5. Furthermore, from [Step 0], the closer the amount of production is to the
best response, the larger the firm A's profit is. Combining these observations, we can see that
producing 1.5 is strictly better for the firm A than producing qA>1.5, for any given production level of
the firm B. In other words, qA>1.5 is strictly dominated by qA=1.5. Hence, the firm A never
chooses qA>1.5.

[Step 2] Consider the second condition: "The firm A is rational and the firm A knows that
the firm B is rational."

Next, suppose that the firm A knows that the firm B is rational. [Step 1] (if applied to the firm B
instead of A) shows that the firm B never chooses qB>1.5 as long as the firm B itself is rational.
Since the firm A knows that the firm B is rational, the firm A knows this fact, too. Given this, qA
=0.75 strictly dominates qA<0.75, as the figure below shows. Thus, the firm A chooses the strategy
neither qA<0.75 nor qA>1.5. From this, we can directly obtain the correct answer for Question 3.

[Step 3] Extend the analysis to further iterations

In the same manner, the firm B also chooses the strategy neither qB<0.75 nor qB>1.5. Hence, it
suffices to consider a situation such that each firm chooses the amount of production
from [0.75,1.5]. In other words, it suffices to consider the green box in the figure below. This green
box is quite similar with the original situation. Thus, we can again repeat the deletion of strictly
dominated strategies in the same way. Then, the green box is going to shrink, and finally converges
into one point (qA,qB)=(1,1), which is equal to the Nash equilibrium. That is, if the rationality of the
firms is common knowledge, they never choose any production levels other than 1. In other words,
in such a situation, they must produce (qA,qB)=(1,1), which is equal to the Nash equilibrium.
4-1 Group Rationality and The Rationality of Individuals
So this situation is described by inefficiency of Nash equilibrium.
So this a terminology frequently used by economists.

And in computer science, the same situation is often described by what is called price of anarchy.
The price of anarchy is the measure of inefficiency of Nash equilibrium, gap between what is rational
for the society and what is rational for each individual. Formally, it's a ratio of maximum total payoff
divided by total payoff at Nash equilibrium.

is always larger than 1

4-2 Why is Group Rationality Different From Rationality of Individuals?

Why Nash Equilibrium is often inefficient?

Nash equilibrium is often inefficient (i.e., it is often not good for players). In such a case, players do not engage in
mutually beneficial (or “good”) behavior.
Which of the following explains this fact?

Good behavior is not publicly observed.

Good behavior is not appreciated by other players.

The benefit of a player’s good behavior often goes partly to other players.

Players compete with each other and try to minimize each other’s payoffs.
4-3 Group Rationality vs. Rationality of Individuals in Biological
Evolution

this = rationality

if species A has maximum fitness, greatest number of offsprings, then the society is eventually
dominated by species A. So this is a very simple idea of the survival of the fittest or
natural selection

This is the very idea of the survival of the fittest, Darwinian idea, but game theory shows that well,

this is not quite right.

So the population of D is increasing, relative population of D is increasing, increasing and increasing
and finally D dominates the whole society, and D is not producing so many offsprings.

therefore game theory has fundamentally

changed the way we view biological evolution. The destination of biological evolution is a Nash
equilibrium played by a game played Nash equilibrium of a game played by selfish genes.
It is a version of Prisoner’s Dilemma.

4-4 Group Rationality vs. Rationality of Individuals in Social Thought

 x-axis =
number of textbooks
in competitive market:
But the best outcome for the society is quite often not Nash equilibrium and
therefore, individuals may deviate from agreement under Laissez-faire.

(Laissez-faire)

4-5 How to Enforce Socially Desirable Outcomes

There are three ways.

Our question is how to sustain, this good point in the society.

First, designing the system or institution or market trading mechanism optimally,

so that the Nash equilibrium is efficient in the first place.
it depends on the situation, but there may be no legal system to enforce a contract.
A prominent example is the one we have seen, the global warming. We don't have international
court powerful enough to enforce a contract signed by countries/
Adam Smith argued that overall social benefit is maximized if buyers and sellers rationally pursue
their own interests in competitive markets. This observation was once over-generalized as follows:

Suppose individuals are in an “unsatisfactory” situation where the social benefit is not maximized, in
the sense that it is possible to make everyone happier. If they are rational, they must (i) recognize
that possibility and (ii) agree to make everyone happier. Therefore, individuals are never trapped in
such an “unsatisfactory” situation, if they are fully rational.

From the point of view of game theory, what is the main drawback of this naïve social thought?

Individuals are often not fully rational.

Even if they agree to make themselves happier, the agreement may be difficult to enforce.
Adam Smith’s idea can be explained by Nash equilibrium.
Even though individuals are fully rational, they may not recognize the possibility of making
themselves happier.
Correct

4-6 Cooperation of gas Stations in Long-Term Relationship

Part I: Need For Cooperation
Not Nash Equilibrium
 Nash equilibrium

Why? How?

They are located side-by-side, and they play this game of price competition again and again, today,
tomorrow, the day after tomorrow. In reality, they play the price competition game every day and
that's what I call long-term relationship. And in the long-term relationship a positive profit may be
achieved.

How to formalize this long term relationship

So the whole dynamic game played over time, today,
tomorrow, the day after tomorrow, this dynamic
interaction is called repeated game.
And price competition game is played every day.
The game played every day is called the stage game of a
repeated game.
 4-7 Cooperation of Gas Stations in Long-Term Relationship
Part II: Mechanism of Cooperation

Trigger Strategy
 Start with cooperation, and keep on cooperating if no one has deviated. (Cooperation is a point that
is better than stage game Nash equilibrium for all players.)

 If anyone deviates, play Nash equilibrium of the stage game forever. (Stage game is the game
which agents play every day.)
Conclusion:
Why people discount future payoff with d?
In the lecture 4.7, we see that future utils are discounted with a rate d, where 0 <d<1. This article
shows what d reflects. We regard that d represents most people's tendency: $1 today is better than
$1 tomorrow. So, why do people tend to prefer $1 today?

Some of the reasons or interpretations: we can regard d as ...

 a psychological factor of impatience. If you are impatient, $1 today is better than $1 tomorrow.
 a measure of uncertainty. In reality, repeated interactions may end by accident, and you may
not be able to obtain $1 tomorrow. Hence, $1 today, which you can surely obtain, is better than
$1 tomorrow.
 a discount rate determined by the interest rate. If you deposit $1 in a bank, it becomes
1+r dollar tomorrow, where r is the daily interest rate. So $1 today has more value than $1
tomorrow.
By these reasons, most people prefer today's 1 dollar.

4-8 Reputation and Brand Name

Restaurant, expensive clothes, car, house, doctor for surgery, lawyer for legal consulting, quality is-
quality of service or goods matters a lot.
But you cannot verify the quality before buying in those, all those instances, and therefore you
may be cheated, 'kay?
But, high quality is assured and the producers or sellers can maintain high profit by means of
reputation or brand name. Sellers with good reputations or brand name can earn extra profit.
And this is a valuable asset they can keep if they keep on producing high quality.
This is how high quality is maintained in the society.
As you can see in those examples, usually there are good company with good reputation
whose quality is high, and they are earning high profit.
And companies with low reputation or no reputation, or companies without any brand name,
supplying low quality product and earning low profit.
And this phenomenon can wonderfully explained by game theoretic model.
4-9 Cooperation in Loosely Knit Organization

In long term relationship, good behavior is reciprocated by the same person.

But in overlapping situation, a good behavior to retiring player is going to be rewarded by future
generation players. So this possibility of cooperation was pointed out by a number of researchers
Cramer, Smith, and including myself.That's the theoretical possibility
 mainly
cooperation happens in protesting activities, attending court session, and attending the labor relation
commission meetings
This is silly and actually we have no evidence that they are playing, trigger strategy.
Тhis is theoretically possible but this is not the mechanism in the union.

one interpretation
another interpretation

Here is the link to the paper:

Kandori, M., & Obayashi, S. (2014). Labor union members play an OLG repeated
game. Proceedings of the National Academy of Sciences,111(Supplement 3), 10802-
10809. http://www.pnas.org/content/111/Supplement_3/10802
4-10 Summary of the Course
Starting with Nash equilibrium and comparing it with reality and, you try to fill the gap.
This is much, much more fruitful way of examining this basic question about, what determines the
two parties' policy rather starting from scratch without any theoretical framework.
Recall the repeated price competition game of two nearby gas stations in the lecture. As discussed
in the lecture, assume that if they cooperate, they want to charge $3, and their cost is $2.
In the above-mentioned game, which of the following best describes the trigger strategy?
Begin with charging $3. If no one deviates, continue to charge $3. If one deviates, charge $2
forever.
Flip a fair coin. If the coin lands heads up, charge $3, and if the coin lands heads down, charge
$2.
If the opponent had charged $3 yesterday, then charge $3 today, and if the opponent had
charged $2 yesterday, then set the price equal to $2 today.
Begin with charging $3. If no one deviates, continue to charge $3. If the opponent reduces the
price, charge $3 forever.

TOTAL POINTS 7
1.Question 1
Consider two gas stations in nearby locations. They face difficulties in sustaining high prices to enjoy
high profits. What is the main reason for that?
1 point
Customers demand that the gas price be decreased over time.
Gas stations cannot agree on the level of price that maximizes their total profit.
Each gas station has an incentive to undercut the high price to steal customers from their
opponent.
High prices do not attract enough customers.

2.Question 2
(Continued from question 1)
If the gas stations could sign a legally binding contract, how would that help them to sustain high
gas prices?
The contract helps them to agree on the right price.
The contract specifies a sufficiently large penalty when a gas station deviates from the high
price specified in the contract.
Gas stations now have moral obligation to sustain high prices.
Customers understand that the gas prices are legally justified.

3.Question 3
(Continued from question 1)
In reality, writing a contract to maintain high prices is prohibited in many countries (because it is anti-
competitive and harmful). When this is true, the gas stations can use their long-term relationship to
sustain their high prices.
What is the basic mechanism of "long-term relationship"?
1 point
In a long-term relationship, players’ hostility towards each other is mitigated.
 In a long-term relationship, players care about the stability of the outcomes.
By deviating from the high price, a gas station can gain some profit today, but its future profits may
decrease.
In a long-term relationship, players can agree on what is the right thing to do.

4.Question 4
(Continued from question 1)
Suppose that, in the absence of a legal contract or a long-term relationship, the gas stations charge
prices equal to cost (=$1) and that they earn zero profit (i.e., assume that this is the Nash equilibrium
when they interact only once).
What is the trigger strategy to sustain a higher gas price = $2?
1 point
Start with price = $2. Then copy the price of your opponent in the following period.
Charge price = $2 for as long as no one has deviated. If anyone deviates, then charge $1 forever.
Always charge price = $2 every day.
Charge price = $2 for the first 100 days and then switch to price = $1 forever.

5.Question 5
Recall the repeated price competition game of two nearby gas stations in the lecture. How does the
long-term relationship work as a mechanism to sustain a favorable outcome? Why do they
cooperate in the equilibrium?
1 point
Because each gas station believes that even if it reduces the price slightly, its sales will not
increase sufficiently, since consumers do not consider such minor changes in price.
Because each gas station believes that, if it reduces its price today, the opponent will also
reduce its price simultaneously. Thus, deviating from maintaining a high price is not at all profitable.
Because each gas station has to pay a large amount of fine to the court, if it breaks the
agreement on maintaining a high price.
Because each gas station finds that the loss from reducing its price today is greater than the
gain from doing so. In particular, if it reduces its price, it will eventually lose future payoffs that could
be earned by maintaining a high price, and this loss is greater than today’s additional payoff
obtained by reducing the price.

6.Question 6
Assume that you share an apartment with your friend. There is a living room in your apartment, and
somebody has to clean it up to keep the room tidy. Every evening, you and your roommate have two
choices: to clean up the room or not to do so. If you choose the first option, you have to spend time
and you do not like this. When both you and your roommate clean up the room, it takes less time
than doing it on your own. Of course, you and your roommate feel happier when the room is tidy,
than when it is messy. However, you prefer to leave the room messy than cleaning it up alone.
Therefore, you and your roommate’s payoff of “cleaning up the living room game,” which you play
every evening, are given by the following payoff table.
As you can see, regardless of what your roommate does, not cleaning up the room is the best option
for you. Consequently, if you and your roommate play this game only for one evening, no one cleans
up the living room, and this is worse for both of you.

In the lecture, it was shown that a long-term relationship can help agents cooperate, and the trigger
strategy was introduced to clarify the point. Which of the following describes the trigger strategy in
this repeated “cleaning up the living room game”?

Clean up the living room every day.

Clean up the living room from Monday to Thursday, and do not clean from Friday to Sunday.
Repeat this pattern every week.
Clean up the living room, if nobody has ever failed to do so. If anyone ever fails to clean up the
room, then you will never do so.
Clean up the living room today, if your roommate did it yesterday. Do not clean up the room, if
your roommate did not clean the room yesterday.

7.Question 7
Assume that your roommate chooses the following strategy; clean up the living room, if no one has
failed to clean up the room; and never clean up the room forever, if either of you did not clean up the
room up. Moreover, assume that you discount the future payoff by d. Then, if you continue to
cooperate with your roommate, in other words, if you continue to clean up the living room every
evening, your discounted sum of payoffs is given by
2+2d+2d^2+2d^3...
What happens if you stop cooperating with your roommate, in other words, you choose “do not clean
up the living room” every evening from today? Which of the following gives your sum of discounted
payoffs?

3+d+d^2+d^3+⋯
3+2d+2d^2+2d^3+⋯
3+3d+3d^2+3d^3+⋯
3

Group rationality” or “what is good for the society (i.e., for players)” can be captured by the notion
of efficient outcomes.
What is the definition of an efficient outcome?

We say that we can “make society better” if we can increase at least one player’s payoff without
decreasing any other players’ payoffs. An outcome is efficient if it is no longer possible to make
society better in that sense.
An outcome is efficient if it is achieved in the most economical way.
An outcome is efficient if the sum of all players’ payoffs is maximized.
An outcome is efficient if it can be achieved by the most advanced technology.

4.Question 4
A restaurant in a touristic location may have an incentive to cheat its customers by providing low-
quality food. How does the “reputation mechanism” explained in the lecture resolve this problem?

A rumour, internet...
Module 4 Optional Challenge Problems TOTAL POINTS 4
1.Question 1
Consider the game in Lecture 2-3, where we had considered that there are N firms and each firm chooses
the amount of production simultaneously. The rest of the conditions are the same as described in Lecture 2-
3.

Now, this static game is infinitely repeated. We examine whether the firms can sustain the amount of
production needed to maximize their joint profit by using the trigger strategy. The discount factor is
d∈[0,1]. Assume that N=2, a=1, b=1, and c=0.
Under this condition, derive dˉ∈[0,1], such that the firms can sustain the amount of production maximizing
their joint profit by using the trigger strategy if and only if d≥dˉ.
Enter the value of dˉ in the form of a fraction, in the answer column below.
1 point
9/17

2.Question 2
Consider the following two-player game, Game 1.

Suppose that this static game is repeated twice (there are two periods). The discount factor is d∈[0,1]. In
the following questions, the strategies constitute an equilibrium if:
-each player chooses the best response to the other player’s strategy in period 2, and
-each player has no deviation incentive in period 1, given that each player chooses the best
response to the other player’s strategy in period 2.
Now let us examine whether both players choose (A,A) in period 1, given that both players choose the best
responses in period 2.
First, find out the combination of the players' actions that is/are the best response(s) to the other in period 2.
Select all the correct statements from the options below.
Both players choose A in period 2.Then, each player is choosing the best response to the other in period 2.

Both players choose B in period 2. Then, each player is choosing the best response to the other in period 2.

Player 1 chooses A and player 2 chooses B in period 2. Then, each player is choosing the best response to the
other in period 2.

Player 1 chooses B and player 2 chooses A in period 2. Then, each player is choosing the best response to the
other in period 2.
3.Question 3
(Continued from Question 2)
Given the pair of actions that is the best response to the other in period 2, derive the equilibrium. Select all
the correct statements from the options below.
There is an equilibrium in which (A,A) is chosen in period 1 when d is sufficiently close to one.
There is an equilibrium in which (B,A) is chosen in period 1 for some d.
For any d, there is no equilibrium in which (A,A) is chosen in period 1.
For any d, there is an equilibrium in (B,B) is chosen in period 1.
(1,2) (2,4) (1,2,4) (1,4) (2,3,4)
3,4
4.Question 4
Consider the following two-player game, Game 2.

As is the case with Question 2, this static game is repeated twice, the discount factor is d∈[0,1], and the
definition of an equilibrium is the same as that of Game 1.
Derive dˉ∈[0,1] such that there is a pure strategy equilibrium in which players choose (A,A) in period 1 if
and only if d≥dˉ.
Enter the value of dˉ in the answer column below.
1 point
1/2