This work focuses on a truce established between criminal groups in El Salvador, commonly known as pandillas. Based on a specific conceptual framework and a repository of methods related to the Actor-Network Theory, I intend to follow some controversies of this process regarding the fulfillment of the agreement, the stabilization of its beginning and end and the determination of its success or failure. To do so, we will follow how the monthly homicide numbers produced by the country s Institute of Legal Medicine participate in these tensions and act organizing different versions of this event. By doing so, we expect to shift our analyses towards others agencies and actors who take part in the life of numbers and, therefore, in the enactment of several versions of the truce. In this way, elements that end up being naturalized and even forgotten become the focus of this work. Thus, this dissertation seeks to expand the body of literature on the truce, while at the same time aiming to broaden the horizons of disciplinary knowledge attentive to the convergence between numbers and situations of violence.