PlayerRank: Leveraging Learning-to-Rank AI for Player Positioning in Cricket

Player prioritization is crucial in sports analysis, yet prioritizing based on playing position is underexplored. This paper focuses on using learning-to-rank machine learning models to select the best players for slots within a cricket team’s batting order in Twenty20 International (T20I) matches. The aim is to build and train position-specific models to rank potential players for each position in the batting order. These models will use listwise ranking algorithms and an artificial neural network architecture to provide data-driven player rankings, enhancing impartiality and performance focus. Each position-specific model is trained to rank players based on their suitability for that position in the batting order, considering factors like performance metrics and specialization. The models are designed to increase impartiality and focus on player performance. The models achieve an average ordered pair accuracy of over 94%, demonstrating their effectiveness in ranking players for specific batting positions. The specialization of positions enhances the utility of the recommendations, providing a more informed approach to player selection. This study highlights the value of using machine learning models to prioritize players based on their suitability for specific batting positions in T20I matches. The models offer an im-partial and performance-focused approach, enhancing the overall quality of player selection in cricket teams.