We argue that standard models of voting do a bad job explaining the frequency of very close mass elections with high turnout. We instead model head-to-head elections as a competition between incentive schemes to turn out voters and elucidate conditions under which parties might prefer close...