Toward Sports Betting as a Financial Asset: An Investigative Analysis of Risk, Investment Potential, and Future Perspectives

Abstract

Sports betting has grown into a large, data-intensive market; however, research on prediction and capital allocation remains fragmented. Most forecasting studies overlook leverage and path-dependent portfolio dynamics, whereas Kelly based frameworks assume frictionless markets and perfectly known probabilities. These gaps have contributed to the failure of several betting funds that developed accurate predictive models but lacked a robust risk-management discipline. To this end, we develop an integrated framework that embeds forecasting, stake sizing, and leverage control into a single portfolio management architecture. First, we model recursive leverage dynamics that generate superlinear growth under favorable outcomes and convex collapse during adverse sequences. Second, we introduce an uncertainty-adjusted Kelly criterion that uses the Kullback–Leibler divergence to penalize model misspecification and reduce stake sizes. Third, we embed them in a three-tier architecture: bet sizing, portfolio CVaR limits and leverage caps. Controlled simulations demonstrate that uncertainty-adjusted staking reduces the ruin probability from 78% to below 2% while preserving 85% of the growth potential, and that VaR-based leverage limits prevent extreme drawdowns (92% vs. 41%). The empirical validation of real-world data collected from the English Premier League confirms that uncertainty quantification improves risk-adjusted performance (Sharpe ratio gains of 30%) and enhances diversification benefits. All results can be reproduced by following the code available at https://github.com/Zabat/betting-risk-management.