I examine the implications of strategic position choices by candidates under two different voting rules: Instant Runoff and Condorcet-Minimax. A neural net forecasts the chances of winning at different ideologies and candidates alter their ideology to maximize their expected utility. This results in different candidate behavior and outcomes than in the non-strategic scenario. Further, the maximal utility strategy differs significantly between the two voting rules being examined. I propose a refinement to the social utility efficiency metric to account for the different utility of the candidate’s chosen positions and use this metric to contrast the outcomes. Finally, this divergence in strategy calls into question the practice of using data (ballots, or survey data) gathered under one voting rule to analyze a different voting rule.