02 - Static Games of Complete Information Flashcards Preview

zGame Theory > 02 - Static Games of Complete Information > Flashcards

Flashcards in 02 - Static Games of Complete Information Deck (21)
Loading flashcards...
1

What is a mixed strategy?

a probability distribution σᵢ = (σᵢ₁, ..., σᵢₖ₁) over Sᵢ
-> each strategy is being played with a certain probability

2

What is a pure strategy?

where σᵢₗ = 1 and σᵢₖ = 0 for all k ≠ l
-> one strategy is played with p = 1, other strategies are not played

3

What is a mixed strategy profile?

σ = (σ₁, ..., σₙ)
-> the set of the set of strategies of each player

4

What is player i's expected payoff from a mixed strategy profile in a normal-form game with finite strategy spaces?

vᵢ (σ₁, ..., σₙ) := the sum of each outcome's utility times that outcome's probability (=each player's probability of playing that specific strategy multiplied)

5

What are possible interpretations of mixed strategies?

- in repeated games, players randomize strategies
- uncertainty about other players
- population shares when randomly choosing individuals

6

What is Theorem 1 in relation to the Existence of Nash Equilibrium?

Any normal-form game with finite strategy spaces has a Nash equilibrium in mixed strategies

7

What is Theorem 2 in relation to the Existence of Nash Equilibrium?

A normal-form game with
- a convex and compact strategy space
- and continuous, quasi-concave payoff function
has a Nash equilibrium in pure strategies

8

What are the assumptions about mixed strategy choices?

- Players’ randomizations are independent,
- Players evaluate mixed strategy profiles by their expected payoffs

9

What is BRᵢ(σ₋ᵢ)?

the set of all best responses against σ₋ᵢ

10

What are the requirements for a mixed strategy profile to be a N.E.?

The mixed strategy profile σ* = (σ*₁, . . . , σ*ₙ) is a Nash equilibrium of G if for each player i:
- vi(σ*ᵢ, σ*₋ᵢ) >= vi(σᵢ, σ*₋ᵢ)
- iff σ*ᵢ ∈ BRᵢ(σ*₋ᵢ)

11

How can you find the mixed strategy equilibrium?

In a mixed N.E., each of the players needs to be indifferent between her pure strategies given the mixed strategy of the other player

12

When is a mixed strategy strictly dominated?

if it assigns positive probability to a strictly dominated pure strategy.

13

Can we lose any mixed-strategy equilibria when eliminating strictly dominated pure-strategies from a game?

NO

14

Can a mixed strategy strictly dominate a pure strategy?

YES, even if that pure strategy is not dominated by any other pure strategy.

15

What does a game being finite tell us about Nash equilibria?

That at least one Nash equilibrium exists - if not in pure, then in mixed strategies.

16

How can you find a mixed strategy that dominates another pure strategy x?

1) give the strategies to be mixed probability p and 1-p
2) set inequalities for each possible strategy the other player could play:
payoffs for mixing p and 1-p > payoffs for x
3) solve the inequalities such that they provide a range for p that dominates x

17

What is the indifference principle?

- in mixed strategies, this principle helps us to maximize probabilities such that a nash equilibrium is found
- indifference = other player has no incentive to switch from either strategy

18

How can you show that there is no pure strategy Nash equilibrium?

1) suppose there is a pure strategy N. E. x̂ = ( x̂₁, x̂₂)
2) go through all cases of relations in strategies and payoffs
3) show that in each, there is an incentive to deviate

19

How to argue that a Nash equilibrium does exist?

1) How does Sᵢ look like? interval = compact + convex
2) is the utility continuous and quasi-concave? (i.e. no weird payoff jumps)
-> if yes + yes, according to Second Existence Theorem, a Nash equilibrium exists

20

How can you find the symmetric Nash equilibrium in a game with i players and a utility function uᵢ?

for each player, uᵢ needs to be maximized (FOC = 0, solve for relevant variable)

21

How to find pareto-optimal (=efficient) symmetric allocation in a game with i players and a utility function uᵢ?

maximize aggregate welfare n•uᵢ (FOC = 0, solve for relevant variable z*)