This paper develops a theoretical framework that formalizes a guerrilla movement's observed oil pipeline attack decision. The movement's decision is modeled as an infinite horizon Markov Decision Process, where each period it chooses to attack or not attack the pipeline. For extortionary behavior, each period the movement's decision will depend on the level of single period payoff and discounted expected future payoffs. We estimate the model using the pipeline attack data and compare parameters when the discount factor is changed. The results suggest that we can reject a zero discount factor hypothesis, demonstrating that the movement's observed attack behavior is compatible with extortionary behavior.