SUPER: Strategies used in public goods experimentation rounds

We asked subjects to submit strategies for playing a specific public good game. The strategies were then made to interact with each other in a computer simulation. afterwards, the subjects twice had the opportunity to revise and resubmit their strategies for a second and a third simulation round. The average contribution to the public good increased from one simulation to the next. While for the first simulation round the dominant strategy solution of the game was submitted 7 times, this strategy had entirely disappeared by the third simulation. The majority of the strategies submitted for the third simulation show the same pattern. In the first period, players contribute the entire token endowment to the public good. Then, in each period, they reciprocate, the other players' average contribution in the previous period. In the final period(s), however, they play the dominant strategy solution, whatever the history of the game.