These lines have a twofold objective: to introduce social simulation as a technique of analysis, and to offer a view of its uses and possibilities. The article starts with a couple of examples aimed to illustrate the practice of simulation. The methodological aspects associated to it are then exposed to, finally, highlight the pros (and cons) of its use. The aim of social simulation is the same as that of statistical equations and ideal types: to generate models of the social reality that help us to build or validate theories about its regularities. What turns social simulation into an innovative analytical strategy is not its objectives but rather its means: programming algorithms that contain the behavioural rules of agents that interact among them and with their environment. The advantages of social simulation are not only experimental but also, and specially, theoretic: it allows the search of the mechanisms that statistical models cannot provide.