Rating Batters in Test Cricket

Abstract

In test cricket, we rated batters’ performance. We have proposed rating criteria as introduced by Scarf, Akhtar, and Rasool in 2014 with additional explanatory variables on the updated data set. The newly added covariates that we included in our research are the home factor and the ground influence. The same rating system is applied in the previous study. Using multinomial logistic regression, sessions from all days of a test match are modeled to determine match outcome probabilities at the end of each session. These models are based on all of the factors that can influence the outcome of a match. It is discovered that the predictors of home factor and pitch quality have a significant impact on the outcome of the test match. We used multinomial logistic regression to model data and estimate the parameters in the models. We forecasted match outcomes using these models at the end of each session and measured batters’ performances by using these probabilities. This process is repeated in a test match at the end of a session, and batters’ contributions to their team score are accumulated. Both teams’ batters are then ranked based on their rating points. The batsmen are rated based on their performance in the match by adding new factors (pitch effect and home advantage) in the models. The proposed ranking is compared with the ICC’s traditional ranking of batters in the test cricket series.