Armies sometimes use fratricidal coercion—violence and intimidation against their own troops—to force reluctant soldiers to fight. How this practice affects battlefield performance remains an open question. We study fratricidal coercion using a mixed-methods strategy, drawing on (1) monthly panel data on Soviet Rifle Divisions in World War II, built from millions of declassified personnel files; (2) paired comparisons of Rifle Divisions at the Battle of Leningrad; and (3) cross-national data on 526 land battles and war outcomes from 75 conflicts (1939–2011) to assess generalizability. We offer three sets of empirical findings. First, coercion keeps some soldiers from fleeing the battlefield, but at the cost of higher casualties and reduced initiative. Second, wartime and prewar coercion (such as mass repression and officer purges) affect soldiers’ behavior in similar, mutually reinforcing ways. Third, the resolve-boosting, initiative-dampening effects of fratricidal coercion generalize across belligerents and wars. Fratricidal coercion generates compliance through fear, compelling soldiers with variable levels of resolve to conform to a uniform standard of battlefield behavior. But the net utility of this approach is dubious. On balance, countries employing fratricidal coercion are less likely to win wars.