In this research paper, we delve into the dynamic field of cricket analytics and match result prediction, leveraging the power of machine learning to enhance strategic depth and narrative in the sport. Our primary aim is to predict cricket match outcomes during the 2nd innings, considering factors such as target, runs left, wickets fallen, and player-specific performance metrics. The models employed in our study encompass Random Forest, SVM Classifier, Logistic Regression, and Naive Bayes. A key innovation in our approach involves the formulation of a custom formula termed 'Player Consistency,' integrating traditional cricket statistics with dynamic player ratings. This novel metric captures the nuanced aspects of player performance, contributing significantly to the predictive accuracy of our models. Our results showcase the superiority of tree-based models, particularly gradient boosted decision trees. The Random Forest model stands out with an impressive testing accuracy of 89.82%, outperforming probabilistic and statistical models. As a comprehensive review article, this paper not only identifies gaps in previous research but also highlights unexplored territories within the realm of cricket match prediction. By providing guidance for upcoming sports analytics machine learning applications, our work aims to contribute to the evolving landscape of cricket analytics, offering valuable insights for both enthusiasts and professional teams.

Cricket Machine Learning Prediction Hierarchical features Comprehensive Dataset selection