Philipp Külpmann speaks at the 10th International Conference of the French Association of Experimental Economics on "On the Predictive Power of Theories of One-Shot Play" on June 20.

Abstract:
We aim to test the predictive power of theories of one-shot (subjects playing a game exactly
once) play in games. We consider a variety of such theories and fix their parameter estimates
from the recent large scale meta-analysis of Wright and Leyton-Brown (2017). Any so calibrated
theory can be identified by the vector of predictions for a set of symmetric hawk-dove games.
We run lab experiments with these games and compare the theories against each other in terms
of their log-likelihood values and by means of Vuong (1989) tests based on these values. Our
main results are as follows. First, and unsurprisingly, all theories have to be rejected at a high
level of significance: no theory explains the data very well. Second, we find that no theory is
uniformly better than all others across all treatments. Third, the theory that provides the highest
log-likelihood for our data is Nash equilibrium with risk aversion. In particular, it beats the two
“winning” theories in Wright and Leyton-Brown (2017).