I develop a model of regime change to study how a revolutionary vanguard may use violence to coordinate and mobilize members of a mass public by convincing them that the level of anti-government sentiment in society is high. The model is consistent with unexpectedly successful violence by vanguards sparking revolution, but also suggests that the relationship between vanguards and the micro-foundations of revolution is complicated and subtle. The model has multiple equilibria, some where revolution is relatively likely and some where it is relative unlikely or impossible. Within an equilibrium, structural factors affect the likelihood of revolution. Nonetheless, if two identical societies play different equilibria, they have very different likelihoods of experiencing revolution, creating a problem for empirical work seeking to identify the root causes of political violence. This fact also suggests that standard empirical arguments about the importance of revolutionary vanguards may be problematic. In equilibrium, there are selection effects—even controlling for all relevant structural factors, vanguards  arise only in those societies that are particularly likely to have successful regime change even in the absence of a vanguard. Finally, on the purely theoretical side, the model studied here is closely related to a global game, yet generates multiple equilibria regardless of informational assumptions. This multiplicity result helps to clarify the substantive content of the technical assumptions underlying uniqueness results in the applied literature on global games of regime change.