A Tournament of Party Decision Rules

Abstract

In the spirit of Axelrod's famous prisoners' dilemma tournaments published in the Journal of Conflict Resolution, we conducted a "tournament of party decision rules" in a dynamic agent-based spatial model of party competition. Entrants submitted rules for selecting party positions in a two-dimensional policy space with unknown voter locations. Each submitted rule was pitted against all others in a suite of very long-running simulations. The most successful rule 1) satisficed rather than maximized in the short run; 2) was "parasitic" on choices made by other successful rules; and 3) used a "secret handshake" to avoid attacking other agents using the same rule. In additional simulations, we show that the same rule wins when we alter the population of strategies and the method by which new parties are assigned strategies. The most successful strategies stayed away from the center of the voter distribution and they tended to make only small changes to their positions between elections.