Philipp Külpmann gives a talk on “Finding the Best Theory for Games with Unique Nash Equilibria in Mixed Strategies” at Bielefeld University on July 6.
Abstract:
In this paper, we experimentally test the performance of different solution concepts using simple games. We are interested in ex-ante theories, i.e., theories which are not calibrated using experimental data from the same experiment. Among others, we compare Nash Equilibria, Quantal Response Equilibria (QRE) and Level-k Reasoning. It turns out that risk aversion is able to explain most of the deviations from Nash in some games and Nash with risk aversion explains the data better than QRE in these games.
