SMU Classification: Restricted

Copyright notice: These materials have been prepared by Nona Pepito to be used exclusively by her students. You
may not reproduce, or allow others to reproduce, adapt, distribute, upload or make available for sale, all lecture
notes and course materials publicly in any way, without your instructor’s written consent. Your instructor holds
copyright to all course materials, including written lectures, power point presentations, study materials, tests and
exercises/projects, pre-recorded video lessons, recordings of lessons conducted online and in the selection of
readings and assignments. The making or distribution of unauthorized copies of course materials is strictly
prohibited and any infringement may be subject to disciplinary action by the University.

Screencasts transcript
Week 2 – Simultaneous-move games (1)

One player does not have a dominant strategy (8 mins. 50 secs.)

This is your instructor, Professor Nona, and in this screencast, we will look at a simple game involving two restaurants.

Imagine two restaurants that operate next to each other. To simplify, assume for the moment that they have identical
food, service, and prices. In such a situation, we would expect customers to be indifferent between the two restaurants.
Suppose that diners come either for Saturday dinner or for Sunday brunch. 70% of the customers come on Saturday,
while 30% come for Sunday brunch. Due to manpower constraints, a restaurant can only open on Saturday night OR for
Sunday brunch, not both.

If both restaurants choose the same day to open, they split the market at that time evenly. If, however, a restaurant is
the only one open at a particular time, then it gets all of the submarket at that time. Space is not a constraint: the
restaurants will always have a table for any customer who wants to dine. Let’s assume that the restaurant managers
don’t talk to each other about their plans. Given all this, the restaurants must decide when to open. What would the
game look like?

Let’s treat this as a simultaneous-move game, so we present this game in normal form. For payoffs, let us use market
shares.

There are two players, Restaurants A and B, and each of these restaurants can open on Saturday or on Sunday. If both
of them open at the same time, they split the market evenly. Since 70% of the customers come on Saturday, this means
that each of the restaurants gets a market share of 35% if both of them open on Saturday. If both of them open on
Sunday, each restaurant gets a market share of 15%: that’s half of the 30% of customers who go on Sunday. Finally, if
they open at different times, then the restaurant that opens on Saturday gets 70% market share, while the restaurant
that opens on Sunday gets 30% market share. Therefore, the normal form (or the payoff matrix) of the game looks like
this: There are two players, each with two possible actions, so we have a 2x2 payoff matrix, with a total of four possible
outcomes.

Let’s get the Nash equilibrium of this game. First, let’s find the best response of Restaurant A. Restaurant A is the row
player, so in any pair of payoffs, we focus on and compare the payoffs that are written first. If A believes that
Restaurant B will open on Saturday, it will also open on Saturday, since 35 is greater than 30. f A believes that
Restaurant B will open on Sunday, it will choose to open on Saturday, since 70 is greater than 15. In short, no matter
what Restaurant B ends up doing, Restaurant A is better off opening on Saturday. Restaurant A HAS a dominant
strategy, which is to open on Saturday.

Next, let’s find the best response of Restaurant B. This time, remember to focus on and compare the payoffs that are
written second. We find that, for B, opening on Saturday is a dominant strategy as well: here, 35 is greater than 30; and
here, 70 is greater than 15.

Both restaurants have a dominant strategy, which is Saturday. This results in a Nash equilibrium where both restaurants
open on Saturday.

Just a word of caution about dominant strategies. If both players have a dominant strategy, then the Nash equilibrium
will consist of actions that correspond to those dominant strategies. However, these dominant strategies may be
different. For example, in the Restaurant Game we saw just now, both players have the same dominant strategy, which
is to open on Saturday. However, there might be situations where the payoffs are such that one restaurant’s dominant
strategy is Sunday, instead of Saturday, in which case the Nash equilibrium would be different.
 SMU Classification: Restricted

Now, suppose Restaurant A implements a loyalty rewards programme. Because of this programme, whenever both
restaurants are open at the same time, Restaurant A gets 60% of the market in that time slot, instead of just half. The
loyalty rewards programme does not affect the payoffs of both restaurants when they choose different time slots. What
will the game look like this time?

Restaurant A’s loyalty rewards programme does not change the payoffs from before if both restaurants choose different
time slots, so these payoffs remain unchanged. This time, however, if both restaurants open on Saturday, A’s loyalty
rewards programme allows it to get 60% of the market share in that time slot instead of just half. 60% of 70% is 42%.
This means that Restaurant B gets the residual, which is 28%. If both restaurants open on Sunday, Restaurant A
captures 60% of the 30% market share in that time slot: that’s 18%. Restaurant B gets the remainder, which is 12%.

Since payoffs have changed, we need to see if the best responses of the restaurants have changed as well. 42 is greater
than 30, and 70 is greater than 18, so it is still the case that Restaurant A’s dominant strategy is to open on Saturday.
For Restaurant B, however, things have changed. If B thinks that A will open on Saturday, it will choose to open on
Sunday instead, since 30 is greater than 28. On the other hand, if B thinks that A will choose Sunday, B is better off
choosing Saturday, since 70 is greater than 12. Therefore, it is no longer the case that B’s best response is always to
open on Saturday, regardless of when Restaurant A opens. Restaurant B no longer has a dominant strategy. This time,
B’s best response is conditional, meaning it depends on what it thinks A will do. Its best response is to avoid competing
with A, to choose the day that Restaurant A does not choose.

A Nash equilibrium is an outcome consisting of actions that are best responses to each other. Because we highlight or
underline a payoff whenever it corresponds to a best response, a Nash equilibrium corresponds to the cell where both
payoffs are highlighted or underlined. When Restaurant A has a loyalty rewards programme, the Nash equilibrium is
Restaurant A opening on Saturday and Restaurant B opening on Sunday. The loyalty rewards programme allows
Restaurant A to capture the bigger submarket.

There are several things to note here:

1. Nash equilibrium does not require both players (or any players at all, as we will see later) to have dominant
strategies. As we saw in the game with Restaurant A’s loyalty rewards, only one player has a dominant strategy,
and yet we have a Nash equilibrium.
2. The first game—the one without loyalty rewards—has a unique Nash equilibrium, meaning it has only one Nash
equilibrium. The second game has a unique Nash equilibrium as well, so the uniqueness of a Nash equilibrium
does not require dominant strategies for both players (or dominant strategies from any player at all, as we will
see later on).
3. Let us focus on the first game (the one without loyalty rewards). Both players have dominant strategies, and the
Nash equilibrium is unique. One might be tempted to think that this game is an example of a Prisoner’s
Dilemma game. However, that would not be correct. In a Prisoner’s Dilemma game, the outcome where they
both choose the other action—the other outcome on the diagonal—has higher payoffs to both players compared
to the Nash equilibrium. That is not the case here: if both restaurants choose Sunday, their payoffs would be
lower.
4. Finally, this is a one-shot, simultaneous-move game that shows that the restaurant who uses a loyalty rewards
programme gains an upper hand: the programme allows it to steal some market share from the other
restaurant whenever they have to compete. But as we see in real life, loyalty rewards schemes are adopted by
many firms. It is rarely the case that a profitable scheme like this is not copied by competitors. Therefore, while
rewarding loyal customers gives Restaurant A an edge here, we would not expect this advantage to last. When
the restaurants interact again, it is likely that Restaurant B will launch its own loyalty rewards programme,
bringing the restaurants back to a level playing field.

No player has a dominant strategy (3 mins. 5 secs)

Game theory screencast

This is your instructor, Professor Nona, and in this screencast, we will look at a simple game involving a bakery and a
coffee kiosk.

In a busy subway station, time is the ultimate currency. Commuters move like a river in the morning rush hour. If a
shop offers a huge menu, the line stalls. To keep up with commuters, our two players—a Bakery and a Coffee Kiosk—
must commit to just one item each morning to ensure they can serve commuters in under 30 seconds.
 SMU Classification: Restricted

So we have two players: the Bakery and the Coffee Kiosk. The Bakery has three types of dough to choose from: its
Budget Scone, its Signature Croissant, or its Luxury Box. The Coffee Kiosk has to choose one of two speed options: its
Basic Drip (which is a quick coffee to make) or its Specialty Latte (which takes longer but goes well with croissants or
luxury pastries). Because the bakery and the coffee kiosk are side-by-side, their profits depend on how well the
commuters think their combined offerings suit people in a hurry.

Let’s define the payoffs in terms of daily profit, measured here in hundreds of dollars. Let’s start with the bakery. The
bakery’s profit from its chosen item depends on the kind of coffee that the coffee kiosk sells. Suppose that if the coffee
kiosk sells its basic drip, the bakery gets the highest profit from selling croissants, followed by scones, then the luxury
boxes. On the other hand, if the coffee kiosk sells its specialty lattes, the bakery gets the highest profit from the scones,
followed by the croissants, then the luxury boxes.

As for the coffee kiosk, if the bakery chooses to sell its basic scones, its profit is higher from the basic drip. The basic
drip is also more profitable for the kiosk if the bakery sells croissants. However, if the bakery sells its luxury pastry
boxes, the kiosk is better off selling lattes.

We now determine the Bakery’s best response. If the kiosk sells its basic drip, its best response is to sell croissants.
However, if the kiosk sells its lattes, the bakery’s best response is to sell scones. This is the Bakery’s best response in
[Link] see that the Bakery does not have a dominant strategy.

As for the coffee kiosk, its best response is to sell its basic drip if the bakery chooses to sell either scones or croissants,
but if the bakery chooses to sell luxury pastry boxes, its best response is to sell lattes. This is the best response of the
coffee kiosk, and we see that, like the bakery, the kiosk also does not have a dominant strategy.

We observe that neither the bakery nor the coffee kiosk has a dominant strategy, and yet a Nash equilibrium exists and
is unique.

– End of transcript –