Cricket Match Prediction Model based on 2023 World Cup Results

Abstract

The objective of this research was to establish a statistical model for forecasting the outcomes of One Day International matches in the ICC Cricket World Cup based solely on data from the match data. Such a predictive model could aid team strategies during intermissions. The study analysed data from 47 World Cup matches from 2023 tournament, excluding those decided by the Duckworth-Lewis method. The primary outcome measured was match results—win or loss. The analysis resulted in the creation of a logistic regression model to predict match outcomes based on weighted predictor variables. Key performance indicators were selected as predictors, including Toss, Opening Partnership Score, Run Scored in powerplay, Wicket lost in powerplay, Total No. of 4’s in an inning and Total No. of 6’s in an inning and Wickets Lost in an Inning. The methodology employed binary logistic regression for predicting outcomes. The research revealed that the logistic regression model developed was significant; however, of all the predictors considered, Total No. of 4’s in an inning and Wickets Lost in an Inning was included in the model. The model was able to correctly predict 82.7% of the match results, which suggests that further refinement, including additional predictors, could enhance the accuracy and reliability of the predictions for future World Cup match outcomes.