IPL Cricket Fantasy Team Prediction for Dream11 using Machine Learning

This study aims to pick the best fantasy cricket team for IPL matches (T20) to participate in fantasy contests on platforms like Dream11 and My11 Circle. This is done by predicting the top eleven players from both the participating teams for a particular match in line with the requirements. A dataset of ball-by-ball details of each IPL match from 2016 to 2023 was obtained from the Cricsheet website [10]. This data was then preprocessed into the required format followed by the addition of important columns such as bowler faced by each batsman and their bowling style. Most importantly we obtained individual batsman-bowler match-ups data in each IPL match along with a fielding dataset as well. We used several Machine Learning models which included regression models such as Random Forest (RF), Extreme Gradient Boosting (XGB) and a baseline Linear Regression Model for the batting predictions, and classification models such as Random Forest, Gradient Boosting, CatBoost (CB), etc for bowling and fielding predictions. Based on the predictions we used a standardized formula to calculate the Dream11 fantasy points scored by the batsman, bowlers and fielders. From the calculated fantasy points, the top 11 players are picked into the fantasy team while the players with the highest and second highest points are selected as captain and vice-captain for bonus points. The model was able to predict at least 7 players correctly as top performers in a match accurately 85% of the time.