This study introduces a theoretical model of how insurgency develops as a function of reactive mobilization. The theory extends a classic distance-decay model by incorporating Kalyvas’ typology of violence. It implies that geographic conditions crucially determine the accuracy of applied violence and thereby its public perception, which in turn determines the actors’ ability to mobilize. As a first test of these effects, I propose a new geographic indicator that expresses the spatial accessibility of a country’s population for both central governments and peripheral insurgent movements. Two empirical implications of the theory are tested with a large-N data set on outcomes and casualties in insurgencies. The new indicator is significantly associated with both military outcomes and the number of casualties in insurgencies since 1970 and strengthens statistical predictions.