Fig 9: Mean payoffs per imitation rule. Maxi-agents are actually the agents winning the most. The mean payoffs decrease with the increase of defecting behaviors in the population.