Observer TFT

As another possibility, gathering information about a future interactant could be efficient as well: observation of a defecting individual would result in playing D when this individual is encountered, reducing the overall payoff to the defector. Consider the strategy Observer tit for tat (OTFT): it is identical to TFT when ignorant of a new partner, but starts by defecting, if this partner was seen defecting on another (Pollock and Dugatkin 1992). On what conditions does OTFT outperform TFT? Remarkably, this only occurs when w is relatively small and the proportion of defectors is sufficiently high. The results of Pollock and Dugatkin (1992) suggest that OTFT only in part compensates for the low value of w, reducing ALLD fitness by diminishing the number of suckers: in a Rover/TFT polymorphism, OTFT is not able to extinguish Rover, it only decreases its frequency (Pollock and Dugatkin 1992). The described features reflect the insight that social controls will only be maintained if exploitation actually poses a severe threat. As the probability of future interaction increases, the chance of punishing defectors increases, rendering third-party observation superfluous.