Experiment 17

Nash Equilibria Normal Form

chance is here not necessary to produce a positive output
- both have a subgame strategy of going to the own venue (top right)
- collusion can only happen when 1 - and only 1 - player is deviating from their subgame strategy to increase the payoffs

analysis
- nothing interesting, perfect 50/50 split as if picked randomly
long-term marriage
- one can think of colluding to always go to either venue together (both opera or both football) which will leave one of the players with a higher payoff
- in reality there will be some arrangement of alternating the venue either each time or occasionally to create equal payoffs for both partners

PD has just 1 clear NE in normal form and 1 better NE if collusion is possible
- as long as I don’t deviate, there is no chance of being hurt by other player
this game has 2 NE where the activity of either player is always hurting the other player through opportunity cost
- risk of a crash which hurts both players considerably

commitment … would be something like being crazy in experiment 5
- strategic move of removing the steering wheel
  - player is now unable to steer away such that the other player is forced to steer to avoid certain destruction
- similar to the doomsday machine which we also watched a video on

crash ratio:
- crash ratio of crashes in random distribution: 25% (1/4)
- crash ratio in experiment: 21.5% → indicates more risk averse than if picked randomly
profit ratio:
- profit ratio in random distribution: 50% (1/4 + 1/4)
- profit ratio in experiment: 52.7% → indicates players which took advantage of risk aversion of other players
overall stay decisions per round (out of 62)
- 1: 30, 2: 31, 3: 28 (symmetric each round)
- about half of individual decisions was to stay
profit per round
- 1: -800, 2: -1800, 3: -1400 (must by symmetric too)
- hypothesis
  - by pure chance in round 2 the non-steering players were matched
  - stay/steer decisions mostly unchanged through all rounds