A new salvo in the punishment wars: if everyone can punish, defectors triumph

arXiv “Punishment can promote defection in group-structured populations”

This paper points out a major problem with theoretical modeling, especially modeling that is simulation-based: tiny changes in assumptions matter. Testing just one set of assumptions takes a lot of effort, and so only through the work of multiple groups do the entire “state space” of possible assumptions get assessed. So when a single paper based on a particular set of assumptions comes out, one needs to temper the impulse to consider the matter settled.