A Novel Regression based Technique for Batsman Evaluation in the Indian Premier League

Abstract

Player profile evaluation and player selection play a very important role in any sports and cricket is not an exception to that. In any cricket team, all players can be segregated into two main roles, namely: batsmen and bowlers. To win a match, a team has to be comprised of the best performing players and such a scenario demands an evaluation of a player profile for team selection. Our study proposes an approach to predict batsman ranking by evaluation of a batsman’s past profile. We have used a supervised machine learning technique on past data of Indian Premier League matches to produce a ranking algorithm of batsmen of an ongoing season, which can be used to gauge a batsman’s performance, or to compare between different batsmen. The novelty of our approach lies in the consideration of multiple cricketing parameters, instead of only one parameter i.e. total runs scored, to design the algorithm and in the prediction of top batsmen.