The rapidly growing popularity of Massively Multiplayer Online Games (MMOGs) has given rise to an increase in the number of players world wide. MMOGs enable many players interact together through a shared sense of presence created by the game. The Peer-to-Peer (P2P) network topology overcomes communication bottleneck problems associated with centralized client/server sys- tems. Thus, P2P-based MMOGs are seen as the way of the future, and many dierent P2P-based MMOG architectures have been proposed to date. However, many architectures are proposed in an ad hoc manner and enhancing the security of such systems is an elusive research problem. In this paper, we address this important issue by making the following contributions. Firstly, we formalize the notion of P2P-based MMOGs and demonstrate that existing P2P-based MMOG architectures can be unied using our model. To our knowledge, this is the rst time that this has been done in the literature. Secondly, we use our model to develop a real-time cheating detection mechanism to identify cheating players, which can be used to expose several MMOG cheating strategies. Finally, we propose a new reputation based system for P2P-based MMOGs to enhance the cheating detection process.