Animation of the Groupish Strategy

The animation below displays a simulation map with shades of green related to agent cooperation in the last 100 interactions. Black cells correspond to complete defection; light green cells correspond to complete cooperation. Gray spots within the central portion of the animation indicate cells where there has been a recent replacement of agent strategy. Simulation parameters: 64x64 map, 30 interactions per generation; 3 with payoff; payoff values of 5, 3, 1, and 0; cooperation error rate of 0.07; communication error rate of 0.75. The map is initially populated at random with agents employing various strategies in equal proportions. Agent strategies include Groupish as well as multiple variations on the following: Tit-for-Tat, Cooperate x%, Firm-but-Fair, Forgiver of x transgressions, and Win-Stay-Lose-Shift. The animation loops infinitely - when it turns all gray, the animation has restarted (generation 0).

Note the early creation of cooperating groups. These are Groupish "cliques". Most of these cliques grow and, over time, merge to form larger cooperating groups. Eventually, they dominate the map. Of course, Groupish dominance is strongly related to simulation parameters.