Monday, March 24, 2008

More evidence that cooperation works--but only in repeated games

In our textbook, we talk about the tradeoff between conflict and cooperation in a prisoners' dilemma. In a repeated game, the best strategies exhibit the following characteristics:
  1. Be nice--no first strikes
  2. Be provokable--retailiate immediately if your rival cheats
  3. Be forgiving
  4. Be clear--make sure your rival can interpret your moves
  5. Dont be envious--focus only on your slice of the profit pie
Now we have more evidence that punishment is not a good strategy.

In Nowak's experiment, the students played more than 8,000 games of prisoner's dilemma, using dimes to reward and punish. The normal game of prisoner's dilemma gives two players two options: cooperate or defect. If both cooperate, each ends up winning a dime. If both defect, each gets nothing. If one cooperates and the other defects, the cooperative player loses 20 cents and the defector wins 30 cents.

Nowak then added a "costly punishment" component. A player could choose to punish someone who didn't cooperate. That penalized the non-cooperative person 40 cents, but the other player had to pay a dime to mete out the punishment.

When Nowak compared how much money people earned or lost in the long run, there was a noticeable correlation between punishment and overall money. The players who punished their opponents the least, or not at all, made the most money.

Those who punished the most made the least money.

1 comment:

  1. I will still not say much on with just these things on the board. I will prefer to stay silent, as it’s not possible to say much. I do trading with world’s finest broker in OctaFX, as they got huge space for me to work out things, as they have lovely demo contest in cTrader, I am able to work things well with been able to do regular practicing that got huge prize up to 400 dollars plus been weekly, it makes things really great.

    ReplyDelete