I work on repeated games, market games and other problems with big, complex and interesting strategy spaces. I use dynamic programming, non-linear optimization, revealed preference and economic experiments.
How to correctly estimate strategies of players by only looking at the results of their repeated interactions?
In this paper we propose an algorithm to reconstruct strategies out of the observed sequence of play in repeated games. The algorithm also accounts for the possibility of measurement and decision making errors, stays agnostic about equilibrium restrictions, and requires only minimal ex ante assumptions. No limited strategy set needs to be assumed.
We show that players use strategies of memory one in experiments, that players do use non-obvious strategies in many games (e.g. Tease-for-tat), and confirm that Australian gas stations learn to collude using day of the week as coordination device.