Sports prediction is becoming popular day by day as a huge amount of data is generated after a single match. The number of methodologies is available for classification of sports data. Machine learning is also one of the best techniques, which obtains good results in sports prediction. In this world of sports, cricket is gaining a huge amount of popularity from the last few decades. The problem of team prediction and winner prediction emerges as one of the challenging tasks in the game of cricket. The winning team combination for the tournament is based on the analysis and evaluation of players from the past study. This type of model builds with several features such as previous record and recent performance of players. The performance analysis of players is a basic problem in every sport including cricket. This performance analysis is used to find the strengths and weaknesses of the players. This work is useful for team management and captain for the selection of players. The prediction of a winner in cricket matches is also one of the complex research problem. Many features are needed for deciding the winner of the match like proper team combination, venue, and weather condition. Cricket is a dynamic game where probability changes with different phases of the game. It is a multi-criteria decision problem. In this paper, the study of some existing methods used for team formation and winner prediction in cricket is done.