Optimizing Horse Racing Predictions through Ensemble Learning and Automated Betting Systems

Horse racing’s burgeoning popularity has spurred extensive research into pre- dictive analysis and strategic decision-making within the sport. The digital age has ushered in an era of unprecedented data complexity and volume related to horse racing, demanding sophisticated analytical approaches. Machine learning, with its broad reach across various sectors, is making significant inroads into the sports betting industry, particularly in horse racing, a domain characterized by substantial financial stakes and widespread interest. Pre- dicting horse racing outcomes is inherently challenging due to the multitude of factors at play. This study endeavors to advance the field by leveraging machine learning algorithms to forecast horse racing results using a compre- hensive dataset encompassing race data from Australia (2015-2024) sourced from Punters.com.au, Australia’s premier horse racing platform, and data from the United Kingdom (including Great Britain and Ireland) spanning 2021 to 2023. A stacked ensemble model incorporating six algorithms – LightGBM, XGBoost, CatBoostRegressor, HistGradientBoostingRegressor, AdaBoostRe- gressor, and TabNetRegressor – was employed to predict the victor of each race. This research can be valuable to stakeholders and researchers alike, offering insights for further analysis and experimentation in this field.

Recommended citation: Nguyen Quoc Khanh, Nguyen Phuoc Sang, Tran Vu Anh, Tran Thi My Quyen (2024), Optimizing Horse Racing Predictions through Ensemble Learning and Automated Betting Systems
Download Paper | Download Slides