Matchmaking in e-sports is the process of pairing players in teams against each other in a fair and competitive manner. Matchmaking systems are crucial in ranked and casual play, ensuring that games are challenging yet enjoyable. So, if done in a centralized manner, the matches can be easily biased, and that's why we propose this project, a matchmaking system using Cartesi Coprocessor.

We implemented the OptMatch, a generalized, iterative, two-stage, data-driven matchmaking framework that requires minimal product knowledge since it only uses match win/lose/score results. The two stages are offline and online. The offline phase prepares the protocol to predict "good" matches, and the online phase effectively does matchmaking. This means that the coprocessor runs the Online phase, and the offline phase is used to build the coprocessor.

One advantage of the chosen framework is that it is applicable to most gaming products. The framework's precision and generic aspect led us to envision this solution as SaaS (Software as a Service), where a game, decentralized or not, can leverage a decentralized matchmaking system.