Halfway handout
-
The whole idea behind game theory:
What happens to me depends on what I do but also on what everyone else does.
What did we do?
- We played the 2/3rds of the average game.
What do I need to be able to do?
- Define an extensive form game and represent a game in extensive form;
- Define an information set.
-
Normal Form Games:
Also called bi-matrix games. This is often used to represent situations in which players make decisions at the same time (and have a finite strategy set).
What did we do?
- We played some matching pennies games (I blogged about the results, links are in the discussion boards).
What do I need to be able to do?
- Define a normal form game;
- Calculate expected utility with mixed strategies;
- Plot expected utilities against/with pure strategies.
-
Dominance:
The idea of being able to ‘remove’ strategies that will never be played.
What did we do?
- Nothing too exciting.
What do I need to be able to do?
- Define dominated strategies (strict and weak);
- Define Common knowledge of rationality;
- Predict rational behaviour using iterated removal of dominated strategies;
- Recognise games that cannot be approached using this tactic.
-
Best responses:
The idea of doing what ‘I’ should do if I know what ‘you’ would do.
What did we do?
- We played a game similar to Golden balls in class. (I blogged about it, link in the discussion board)
What do I need to be able to do?
- Identify best responses in pure strategies (underline stuff);
- Identify best responses against mixed strategies;
- Define the sets \(UD_i\) and \(B_i\);
- State and prove certain cases of theorems connecting these two sets.
-
Nash equilibria in pure strategies:
Here we are officially defining what we meant by ‘predicted rational behaviour’. In essence: ‘not having a reason to change what you are doing’.
What did we do?
- Nothing too exciting.
What do I need to be able to do?
- Give definition of Nash Equilibria in pure strategies;
- Continue to use best responses to identify these;
- Calculate Nash equilibria in pure strategies when strategy sets are continuous.
-
Nash equilibria in mixed strategies:
Congratulations we’re now doing real game theory.
What did we do?
- We played a Rock, Paper, Scissor, Lizard, Spock game in class (I blogged about it, link on a discussion board).
- The weak after seeing this we played against a random number generator (I blogged about that too, link in the same place).
What do I need to be able to do?
- Define Nash equilibrium (completely, ie in mixed strategies);
- Identify Nash equilibrium using best responses against mixed strategies;
- State and prove the Equality of Payoffs theorem;
- Apply the equality of Payoffs theorem;
- State and prove specific cases of Nash’s theorem.
-
Extensive form games (again):
We go back to the tree things and get a bit more precise.
What did we do?
- We played the centipede game but I’ll come back to that when we talk about subgames.
What do I need to be able to do?
- Recognize valid extensive form games;
- Define sequential rationality;
- Define backwards induction;
- Prove the theorem of existence of Nash equilibrium in games of perfect information;
- Solve extensive form games using backwards induction;
-
Subgame perfection:
We connect Normal form games and extensive form games and define a stronger type of equilibrium condition.
What did we do?
- We played the centipede game and analysed the equilibrium strategies: everyone should take AT ALL STAGES.
What do I need to be able to do?
- Write down the normal form representation of an extensive form game;
- Define a subgame;
- Define subgame perfect equilibria;
- Obtain subgame perfect equilibria in games.