Students from the Faculty of Information Technology at CTU in Prague (FIT CTU) placed second in the prestigious Qminers Quant Hackathon. The competition challenged participants to develop predictive models and betting strategies using real data from ice hockey matches and to try to outperform a virtual bookmaker. The hackathon attracted 250 participants from across the Czech Republic.
The AlphaPrime team from the Faculty of Information Technology at CTU in Prague, consisting of Ondřej Skácel and Patrik Herman, achieved second place and won a prize of CZK 30,000. Participants in the hackathon took on the role of professional sports bettors, whose goal was to earn as much virtual money as possible. The principle of sports betting is similar to investing in financial markets – it is about finding a mismatch between price and reality and investing where an opportunity is perceived. The competition was not about knowledge of a specific sport (in this case, ice hockey), but about the ability to translate it into an effective computer algorithm. As on the stock market, the winner was the one who maximized profit and minimized risk.
The participants were given real data from ice hockey matches, to which they applied machine learning methods and created a betting strategy that outperformed a virtual bookmaker. The task consisted of two main steps: building a predictive model of match outcomes and designing an investment strategy with respect to the bookmaker’s odds – that is, deciding what to bet on and how much.
“The qualification round focused on understanding the data and building a reliable model foundation, while the final took place in person and allowed only a single submission. Thanks to the solution prepared during the qualification round, we only fine-tuned the details,” says Ondřej Skácel. “Our model was based on logistic regression, which predicted a home team win based on Glicko ratings of individual match statistics – this can be interpreted as the team’s strength in a given aspect of the game,” he explains.
Qminers Quant Hackathon 2025 took place on December 6 in Prague and was open to all students. The event was organized by Qminers, a company specializing in algorithmic trading, in cooperation with the IDA (Intelligent Data Analysis) research group from the Faculty of Electrical Engineering at CTU in Prague.
“Most teams opted for simpler models, mainly logistic regression or decision trees. However, we also saw more advanced feature engineering or Kalman filters. Correctly handling risk was critical: how much to bet depending on how much money you have, how confident you are, what is allowed, and what changes during the competition. Those who stuck with fixed estimates often ran into trouble,” says Tonda Kříž, a Python developer at Qminers.
