B Wiedenbeck and MP Wellman
Multiagent simulation extends the reach of game-theoretic analysis to scenarios where payoff functions can be computed from implemented agent strategies. However this approach is limited by the exponential growth in game size relative to the number of agents. Player reductions allow us to construct games with a small number of players that approximate very large symmetric games. We introduce deviation-preserving reduction, which generalizes and improves on existing methods by combining sensitivity to unilateral deviation with granular subsampling of the profile space. We evaluate our method on several classes of random games and show that deviation-preserving reduction performs better than prior methods at approximating full-game equilibria.