CRICKT20 – Squad machine learning approach to predicting a squad

Abstract

" Cricket is a sport where teams of less than a dozen players compete against each other in the form of a match. Players in each team can be categorized based on their role in the field. Despite eleven players being on the field, an additional four players are added to the squad so that players can be swapped out in case of injuries or less than stellar form. These fifteen players in the squad are selected from a pool of players. The process involved in manual, involving selectors, coaching staff and the Captain. Due to the process being manual, several factors can affect the selection process such as personal biases and political influences which turns the process into subjective judgement instead of an objective judgement. This can result in better teams being snubbed of their opportunities in succeeding due to less than stellar players being selected. This process requires the elimination of personal judgement and a shift into a more objective process. Thus, the author proposes a solution for T20 Squad Selection by using Machine Learning Models to recommend squads from pools of players. Previous two years of performance, previous year performance in Premiere Leagues such as IPL, PSL, BPL, LPL, BBL and CPL, Performance in Domestic circuits and Under 19 performance are considered for each player to ensure that youngsters and well-experienced players aren’t disadvantaged in these processes."