Comparing Pay versus Performance of IPL Bowlers: An application of Cluster Analysis

Cricket is one of the sports where a large amount of data is generated in every game. In addition to popularity of existing formats of the game, the grand success of India at the first Twenty20 Cricket World Cup led to creation of the Indian Premier League (IPL) in 2008. It is a franchise based tournament where teams are formed by competitive bidding from a pool of Indian and International players. Since player salaries are determined through auctions, performance of individual players is regularly monitored by media and team owners. The purpose of this ex-post study is to investigate the relationship between player performance and valuation. Along with developing new performance metrics, we applied data mining (K-means cluster analysis) and identified four distinct groups of bowlers based on performance effectiveness. To validate findings, we overlaid cluster results by team ranking in 2013 and observed significant differences in team mix of top and poor performing teams. The study can be applied to assist team owners in deciding players to be retained for the next season, those who should be traded-in and those in need of additional coaching/mentoring.