Game Theory experiment in Second Life

a.k.a how i earned some lindens

on Thursday a group of students from a German university is conducting an experiment in Second Life testing a cooperative game. It goes as follows ( as I relate from my experience):

• You are recruited into the game, you are given 1000L to start with but your earnings depend on the decisions you and the other person make during the game


• You are either person A or person B in the experiment and your role is assigned to you through a random generator. If you are A you stay in the room shown above, otherwise you get transported to one of the sky cabins by selecting teleport on the blue columns. This is basically to make sure that the pair don't see eachother during the game and can't neogotiate in private.


• A starts by offering B a certain amount of money. The "moderator" in the game quadruples the amount and give it to B. I was offered 500L, so I end up receiving 2000L, in total I now have $3000, including my initial 1000L


• Now B offers back an amount to A and the game ends. This amount is not quadrupled and B can offer A 0L if he wants. I offered the A person in my game 1000L.


• end result, I have 2000L and the other person has 1500L

Note that the more A offers B the bigger the pot but also the stake is larger for A because what if B does not respond to A's offer of goodwill and gives back o in return?

The optimum strategy, theoretically, in this case for B would be to keep all the money A gives him since the game ends after his decision. However what people actually decide to give back reflect a preference for "fairness". I for one, thought A is pretty fair by offering me 500 so I gave 1000 back. Had he/she only given me a meager amount, I ( or another player) might be less inclined to cooperate.

I don't know what is the aggregate result yet for this experiment, I'll go and ask. Nevertheless, it's a very novel way of testing as traditionally recruits are usually college students with similar background. Second Life allows for a better cross section although there is the risk of people using double avartars to do the experiment.