For several weeks, FootProbability has been publishing its analyses daily, based on a proprietary statistical model. After a more challenging period, the latest results show a marked improvement in performance, with a series of bets allowing the model to regain positive momentum.
Over the last seven days, FootProbability has posted a return on investment (ROI) of +45.2%, with 23 winning bets out of 36 and a profit of 16.27 units. This improvement contrasts with the overall season results, which remain slightly negative with an ROI of -2.2% after 575 bets. This discrepancy perfectly illustrates the reality of sports betting: performance is evaluated over the long term and naturally experiences periods of both growth and decline.
This positive run is driven by several particularly successful days. On July 10th, all four Top 3 selections were validated. On July 11th, the model notably predicted victories for Turku PS and Ulsan Hyundai, while also validating several Over 2.5 markets. On July 12th, the wins of FC Astana, Incheon United (at odds of 3.83), Ulytau, and Malmö FF (Over 2.5) further confirmed this positive momentum.
These results are all the more interesting because the model continues to search for Value Bets, that is, situations where the calculated probability is higher than the implied probability of the odds offered by bookmakers. The goal is not to predict every match with certainty, but to identify opportunities offering a positive expected return in the long run.
Recent progress also shows that the adjustments made to the model continue to have an impact. Filtering competitions, improving the selection of Value Bets, and optimizing confidence criteria now allow for more selective analyses. The day of July 13, 2026, is a good example: only four clean matches were selected, with just one true Value Bet. Rather than multiplying predictions, FootProbability now prioritizes quality over quantity.
The goal now is to confirm this trend in the coming weeks in order to bring the overall ROI back into positive territory. With an estimated bankroll of €87.53 from a starting bankroll of €100, the model has already made up a significant portion of its ground and continues its upward trajectory.
At FootProbability, transparency is an integral part of the project. All results, whether positive or negative, are published so that everyone can follow the actual evolution of the statistical model. This approach provides a true picture of its performance and highlights what really matters: the ability to generate long-term value rather than seeking short-term gains.