This paper proposes a comprehensive research program for determining the epistemological foundations of analytic wargaming. Wargaming has been used in military, government, and private sectors for decades, with tens of millions of dollars spent annually on it. In light of the changing strategic circumstances of the twenty-first century, it has only become more popular. However, the epistemological foundations of the method are poorly understood. Many professional wargamers contend that wargaming is an “art” and thus unable to be systemically evaluated. Recent work by a small coterie of international relations scholars has contended that wargaming can be reconciled with social science, typically by evaluating wargaming according to experimental standards. However, this solution strips wargames of most of their unique features and cannot explain why some of the most prominent wargames in history produced meaningful results. In this paper, I argue that in the attempt to better understand wargaming’s epistemology, scholars should begin by recognizing the prominent features of wargames and research each of these to determine if and how wargames produce rigorous knowledge. In making this argument, I identify five distinct “methodological machineries” of wargaming—the recurring processes through which wargames may produce knowledge—that distinguish wargaming from other social science methods: (i) they are representative, (ii) they feature consequential decisions made by human players, (iii) they are adjudicated, (iv) they are immersive, and (v) they are bespoke designs. I show how each of these machineries offers potential opportunities and dangers in the production of knowledge through the method of wargaming. In outlining these distinct features, I offer a clear and viable research program for epistemologists of wargaming.