What should major parties out of power do to win elections? To answer that question, we need to understand what these parties do to recapture political ascendancy and whether their actual behaviour differs from their optimal behaviour. In this article, we propose a systematic, replicable method of identifying the competitive strategies that American parties out of power have adopted in their pursuit of the presidency. We present a taxonomy of party strategies, which we operationalize by comparison of utility functions for hypothetical voters. Using both directional and proximity models of issue voting, we compute these utility functions for each presidential election from 1852 to 1996, controlling for variables that systematically affect voting, including economic conditions and incumbency. These results suggest that, contrary to the views of many political scientists and party activists, there is no single optimal strategy through which parties out of power can regain it. Rather, several competitive strategies offer similar prospects for electoral success.