Evaluating profitability in sports betting using probabilistic models and betting strategies

Pál, József Gergő and Bíró, Csaba (2025) Evaluating profitability in sports betting using probabilistic models and betting strategies. ANNALES MATHEMATICAE ET INFORMATICAE, 61. pp. 202-214. ISSN 1787-6117

Preview

Text 202_214_pál.pdf - Published Version Download (7MB) | Preview

Abstract

Sports betting has evolved into a multibillion-dollar global industry, raising the question of whether consumers can achieve sustainable long-term profit. This study explores whether combining probabilistic prediction models with various betting strategies can yield statistically significant profit in football betting. We examine six prediction models – including Poisson, logistic regression, Elo, Monte Carlo simulation, and two novel heuristics (Veto and Balance) – alongside five popular betting strategies: Flat Betting, Martingale, Fibonacci, Value Betting, and the Kelly criterion. A custom Python-based simulation system was developed using real match data from 539 unique football games played between March and May 2025. A total of 10,000 match groups were generated, each containing 25 unique matches, yielding 300,000 model–strategy runs (6 × 5 × 10,000). Simulations preserved chronological order and modeled realistic stake adjustments. Our results highlight the complex relationship between predictive accuracy and profitability, and the limitations of exploiting statistical advantages in an efficient market. While some combinations showed short-term gains, consistent long-term profit remained elusive under most conditions. The findings provide insight into model performance, risk management, and the practical challenges of algorithmic sports betting. This study is intended solely for academic purposes; the results should not be interpreted as practical betting advice.

Actions (login required)

Edit Item