Context Based Cricket Player Evaluation Using Statistical Analysis
Published: 2019
Author(s) Name: Vaibhav Khatavkar and Parag Kulkarni |
Author(s) Affiliation: Research Scholar and Assist. Prof., Dept. of Comp. Engg. and Informn. Tech., College of Engg., Pune
Locked
Subscribed
Available for All
Abstract
In business, making decisions on the spot is a critical task since data is available online. Same applies for sports. For managers and coaches selection of best fit player in the team has to be made. Many times, team managers have to bid for the players. Many statistical methods are applied in order to evaluate a player for sports, specially, baseball, and football. This paper focuses on Cricket since it has been deprived of good statistical methods. Typically, for cricket, Deep Performance Index (DPI) method is used to evaluate a player. But it has some limitations. This work aims to assimilate the best features of the previous approaches and losing the drawbacks. An algorithm is devised for statistical analysis of cricket data and particularly Indian Premier League (IPL) data which can be later extended to other sports with minor customizations. The algorithm considers context of the player in terms of indices like most valuable player index for batsman and for bowler. The important context introduced is winning contribution ratio of the player. The algorithm is tested on IPL and proves to be comparable with DPI. Hypothesis testing is performed to validate the results.
Keywords: Context, Cricket data mining, Hypothesis testing, Indian Premier League (IPL), Statistical analysis.
View PDF