The possibility that actors strategically condition their behavior onpartially unobservable factors poses a grave challenge to causal inference,particularly if only some of the actors whose behavior we analyze are atrisk of experiencing the outcome of interest. We present a crisis bargainingmodel that indicates that targets can generally prevent war by arming. Wethen create a simulated data set where the model is assumed to perfectlydescribe interactions for those states engaged in crisis bargaining, whichwe assume most pairs of states are not. We further assumeresearchers cannot observe which states are engaged in crisis bargaining,although observable variables might serve as proxies. We demonstrate that anaïve design would falsely (and unsurprisingly) indicate a positiverelationship between arming and war. More importantly, we then evaluate theperformance of matching, instrumental variables, and statistical backwardsinduction. The latter two show some promise, but matching fares poorly.