Optimal Choice Problems Strategic Interdependence
Optimal Choice Problems Strategic Interdependence
1. Introduction
Introduction
• Game theory as a topic in Math:
• How to formulate and solve optimal choice problems when there is strategic
interdependence? Example: The cold war.
• Strategic interdependence
• Elections, oligopoly, lobby, war, corporate finance, regulation, IO,...
1
15/07/2019
1. What is a game?
• The actions that are best for an individual to take may depend on what he
expects the other players to do
Basic elements
• The Players: interactive decision-makers.
• The Rules:
• Who moves when?
• What do they know when they move?
• What can they move?
• The Outcomes: what players care about.
• The Payoffs: representation of players’ preferences over outcomes
2
15/07/2019
Classification
1. Time
• Static games (simultaneous-move) x Dynamic games (sequential move)
2. Information
• Complete information x Incomplete information
3
15/07/2019
2. Extensive-form games
• Captures who moves when, what action each player can take, what
players know when they move, what the outcome is as a function of
the actions taken by the players, and the player’s payoffs from each
possible outcome.
• Example: modified version of matching pennies
4
15/07/2019
Information set
• Information set: it is a subset of a particular player´s decision nodes
• When play has reached one of the decision nodes in the information set and
it is that player´s turn to move, she does (or does not) know which of these
nodes she is actually at !
• Example: Matching Pennies
5
15/07/2019
Note
• Action set (A): set consisting of all the actions that, under some
circumstances, are available to the player.
6
15/07/2019
Why?
• Player 1: {H, T}
• Player 2:
• Example
• The prisoners' dilemma
7
15/07/2019
• The police explain the consequences that will follow from the actions
they could take
• If both confess then both will be sentenced to jail for six months
• If one confesses but the other does not, then the confessor will be released
immediately but the other will be sentenced to nine months in jail
• Six for crime
• Three for obstructing justice
8
15/07/2019
• Each player has two strategies: Confess (or fink) or Not confess (or be
mum)
• We implicitly assume that each player does not like to stay in jail
9
15/07/2019
𝑢: 𝑆 → [−∞, +∞]
∈
𝑠 → 𝑢 (𝑠)
where 𝑢 (𝑠) is the payoff of player i when he plays strategy 𝑠 and any other
player j plays strategy 𝑠
𝑢 𝑠 = 𝑢 𝑠 ,𝑠 ,…,𝑠 ,𝑠 ,𝑠 ,…,𝑠 = 𝑢 𝑠 ,𝑠
10
15/07/2019
4. Questions
11
15/07/2019
1. Simultaneous move
• A normal form game (or strategic form) is usually related to simultaneous
move games.
• As we already know, this does not imply that they act simultaneously
• It suffices that each player chooses his or her action without knowledge of the
others' choices
• For example, in the prisoners' dilemma, the prisoners may reach decisions
at arbitrary times, but it must be in separate cells.
• Notice, however, that we are also able to represent extensive games in
strategic form (later…)
∀𝑠 ∈ 𝑆 , 𝑢 𝑠 ,𝑠 < 𝑢 𝑠 ,𝑠
Rational players do not play strictly dominated strategies, regardless of the strategies
that he anticipates his rival will play!
12
15/07/2019
• This results in a worse payoff for both players than would (Not confess, Not confess)
13
15/07/2019
2. Each step requires a further assumption about what the players know about
each other's rationality
3. To apply the process for an arbitrary number of steps, we need to assume that
it is common knowledge that players are rational
4. This process often produces a very imprecise prediction about the play of the
game
14
15/07/2019
• The process produces no prediction whatsoever about the play of the game
• Question
• Is there a stronger solution concept than iterated elimination of strictly dominated strategies
but which produces much tighter predictions in a very broad class of games?
• In order for this prediction to be compatible with incentives (or correct) it is necessary
that each player be willing to choose the strategy predicted by the theory
• Thus each player's predicted strategies must be that player's best response to the
predicted strategies of other players
• Because no single player wants to deviate from his or her predicted strategy
• A solution of the game satisfying the previous property is called a Nash equilibrium
15
15/07/2019
i.e,
∀𝑠 ∈ 𝑆 , 𝑢 (𝑠 , 𝑠 ∗ ) ≤ 𝑢 (𝑠 ∗ , 𝑠 ∗ )
That is, for each player i, 𝑠 ∗ is player i’s best response to the strategies specified for the n-1
other players
16
15/07/2019
2)
3)
17
15/07/2019
• While at workplaces, Pat and Chris must choose to attend either the
opera or a rock concert
• Both players would rather spend the evening together than apart
• We will see that in some games with multiple Nash equilibria one
equilibrium stands out as the compelling solution: in particular a
convention can be developed
• In the example above, the Nash equilibrium concept loses much of its
appeal as a prediction of play since both equilibria seem equally
compelling: none can be developed as a convention
18
15/07/2019
• Each player has a penny and must choose whether to display it with heads or tails facing
up
• If the two pennies match then player 2 wins player 1's penny
19
15/07/2019
4. Mixed strategies
• A mixed strategy is a probability measure (distribution) over the strategies in 𝑆 . A strategy in 𝑆 is
called a pure strategy
∀𝑠 ∈ 𝑆 , 𝑝 = 𝑝 𝑠 ≥ 0 and 𝑝 𝑠 =1
∈
0 𝑖𝑓 𝑠 ≠ 𝑠̂
∀𝑠 ∈ 𝑆 , 𝑝 𝑠 =
1 𝑖𝑓 𝑠 = 𝑠̂
𝑢 𝑠 ,𝑝 ≡𝔼 𝑢 𝑠 = 𝑝 (𝑠 ) 𝑢 𝑠 , 𝑠
∈
( )
• The expected value of agent i's payoff if he plays 𝑝 believing that the other players will play
according to 𝑝 is denoted by
𝑢 𝑝 ≡𝔼 𝑢 = 𝑝 𝑠 𝑢 𝑠 ,𝑠 = 𝑝 𝑠 ×𝑝 𝑠 … × 𝑝 (𝑠 ) 𝑢 𝑠 , 𝑠
∈ ∈
20
15/07/2019
∀𝑖 ∈ 𝐼, 𝑝∗ ∈ 𝑎𝑟𝑔𝑚𝑎𝑥 𝑢 𝑝 , 𝑝∗ : 𝑝 ∈ ∆(𝑆 )
• However, a mixed-strategy Nash equilibrium does not rely on any player flipping coins,
rolling dice, or otherwise choosing a strategy at random
• Player j's mixed strategy is interpreted as a common statement of any other player i's
uncertainty about player j's choice of a pure strategy
• Under this interpretation each player chooses a single action rather than a mixed strategy
• A Nash equilibrium (𝑝∗ ) ∈ in mixed strategy is such that any pure strategy in the support of 𝑝∗ is
a best response given that player i’s beliefs about player j's actions is 𝑝∗
21
15/07/2019
• Suppose that player 1 believes that player 2 will play Heads with probability 𝑞 and
Tails with probability 1 − 𝑞
• Given this belief we have
22
15/07/2019
𝑟𝑞 1 + 𝑟 1 − 𝑞 (−1) + 1 − 𝑟 𝑞 (−1) + 1 − 𝑟 1 − 𝑞 1
= 𝑞 4𝑟 − 2 − (𝑟 + 1)
• Player 2’s best response (𝑞 ∗ 𝑟 ) is:
𝑞 ∗ 𝑟 ∈ 𝑎𝑟𝑔𝑚𝑎𝑥 𝑞 4𝑟 − 2 − (𝑟 + 1)
1
𝑞 = 0, 𝑖𝑓 𝑟 <
2
1
𝑞∗ 𝑟 = 𝑞 = 1, 𝑖𝑓 𝑟 >
2
1
∀𝑞 ∈ [0,1], 𝑖𝑓 𝑟 =
2
23
15/07/2019
24