A key piece of conventional wisdom among scholars of modern armed conflict is that collateral damage is often strategically costly in war. Yet most combatants already know this and take actions after mistakes—most prominently, the distribution of “condolence payments” to civilian victims—in order to mitigate these costs. Do these payments work? This question is important not only for policymakers but also for deeper theoretical debates about how civilians respond to combatant signals in war. To examine these issues, I use micro-level conflict event data on 4,046 condolence payments made by Coalition forces to civilian victims during the Iraq War from 2004 to 2008, matching it with corresponding data on collateral damage and insurgent violence. The results of this analysis reveal that post-harm compensation does significantly diminish local rates of insurgent violence, and that this is true across different types of payments (cash handouts or in-kind assistance). Ultimately, these patterns can be best explained by a rationalist mechanism in which civilians update their beliefs about violent events based on new information about combatants’ wartime intentions. The results thus provide a compelling strategic rationale for combatants to compensate their victims in war, and suggest that civilians are not blinded to new information about conflict dynamics by their preexisting biases.