Week4 - Behavioural Game Theory

• what are simultaneous games also known as and what are its characteristics

  i. players

  ii. strategies

  iii. payoffs

  
  

  also known as normal form games
• what are sequential games also known as and what are its characteristics

  i. players

  ii. strategies

  iii. payoffs

  iv. decisions nodes, terminal nodes, branches, path

  v. information

  
  

  also known as extensive form games
• how would you solve normal form games

  i. pure strategy Nash equilibrium (PSNE)

  a. dominance

  b. iterated dominance

  c. best response analysis

  
  

  ii. Mixed strategy Nash equilibrium (MSNE)

  a. a player will mix iff they are indifferent between strategies
• how would you solve extensive form games

  i. subgame perfect Nash equilibrium (SPNE)

  a. backward induction (perfect information)

  b. solve subgames (imperfect information)
• what are the characteristics of level-K players

  level-K divides players into different cognitive levels

  
  

  L0 - nonstrategic

  L1 - anchors their beliefs on what L0 will do

  L2 - anchors their beliefs on what L1 will do

  
  

  each player thinks their partner is one level below them
• explain what team reasoning is

  its the best outcome for the team - each player does their part for the best solution expecting everyone else to do the same
• explain the centipede game

  i. these are dynamic games of perfect information where two players make alternating moves

  ii. at each move, either D can be played where the game ends or A can be played which passes the move onwards

  iii. the total payoff doubles as one moves across the nodes
• theoretically, what should rational players choose to do in the centipede game

  they should choose D 100% of the time
• explain the quantal response equilibrium

  players play noisy best responses

  
  

  players correctly anticipate the noisy best replies of others

  
  

  i.players will choose the strategy which provides them with higher payoff

  ii. the probability of each strategy is calculated
• in the formula used for quantal response equilibrium, what does lambda represent

  lambda informs us about the degree of noise in the choices

  
  

  as lambda tends to infinity, the variance of the shocks tends to zero (Nash equilibrium)

  
  

  as lambda tends to 0, the probability tends to 0.5 (random choice)