Dynamic behavior and player types in majoritarian multi-battle contests

In a dynamic contest where it is costly to compete, a player who is behind must decide whether to surrender or to keep fighting in the face of bleak odds. We experimentally examine the game theoretic prediction of last stand behavior in a multi-battle contest with a winning prize and losing penalty, as well as the contrasting prediction of surrendering in the corresponding contest with no penalty. We find varied evidence in support of these hypotheses in the aggregated data, but more conclusive evidence when scrutinizing individual player behavior. Players’ realized strategies tend to conform to one of several “types”. We develop a taxonomy to classify player types and study how these types interact and how their incidence varies across treatments. Contrary to the theoretical prediction, escalation is the predominant behavior, but last stand and surrendering behaviors also arise at rates responsive to the importance of losing penalties.