The idea of the selection of the correct players plays a crucial role in the team formation of any sport like it does for Cricket. Enforcing a point system, player points were predicted using historic match data. Different machine learning approaches were integrated and compared to form a model ensemble. Subsequently, using genetic algorithm (GA), these points were used to embody the model fantasy team. Our findings show that various approaches yield different results, and hence, an amalgamation of such approaches can be utilized. The top players from each team could be recognized with as many as three being accurately identified out of five. Additionally, accuracy above 61.2% was secured in the formation of the fantasy team. The paper thus entails such decisive information to all the key stakeholders. Using this approach, one can thus predict, optimize, and in turn build their fantasy team with the least losses.