| 3 Bilder
The 2025 European Women's Football Championship kicks off tomorrow, Wednesday, in Switzerland. The Spanish national team is favored to take home the trophy in the three-week tournament. Germany, France, and England are also strong contenders for winning the European title – according to a team of statisticians from the University of Innsbruck and TU Dortmund University.
The 2025 European Women's Football Championship kicks off tomorrow, Wednesday, in Switzerland. The Spanish national team is favored to take home the trophy in the three-week tournament. Germany, France, and England are also strong contenders for winning the European title – according to a team of statisticians from the University of Innsbruck and TU Dortmund University.
The UEFA Women’s Euro 2025 will take place in Switzerland from July 2 to 27, 2025. According to the current forecast, the Spanish team has a 27.2% probability of winning. The research team, consisting of Andreas Groll, Marjan Farahani and Rouven Michels from TU Dortmund University, Germany, and Achim Zeileis from the University of Innsbruck, Austria, created this prediction using a hybrid machine learning model.
The forecast combines several statistical models in two steps: In the first step, two complex statistical models are used to determine the strengths of all teams and their players based on different information - on the one hand, team strengths were determined based on matches in the past, and on the other hand, a bookmaker model based on betting odds for the expected team strengths in the upcoming tournament. “In the second step, an ensemble machine learning process, a so-called random forest, decides how the two models with their resulting team strength estimates as well as other information about the teams can best be combined,” explains Andreas Groll from TU Dortmund University.
“With this model, we ran simulations to determine the likely course of the tournament,” explains Achim Zeileis from the Department of Statistics at the University of Innsbruck. “Of course, the actual outcome of the European Championship can deviate from this. That's clear to us as statisticians, but that's also what makes the sport so exciting,” smiles the expert, but emphasizes: “The same prediction method would have had correctly predicted England as European champions for the last women's tournament in 2022.”
Spain is ahead
The researchers ran 100,000 simulations of the entire European Championship based on the prediction model - match by match, following the tournament draw and taking all UEFA rules into account. This results in probabilities for each team to progress to the next rounds of the tournament and, ultimately, to win the European Championship. The favorites at this European Championship are Spain with a probability of 27.2 percent, followed by Germany (23.0 percent), France (17.6 percent) and England (17.2 percent). All other teams are well behind.
The simulation shows that Spain in particular has benefited from the tournament draw. “Spain is in the comparatively easy Group B and will face a team from the relatively weak Group A with hosts Switzerland if they progress to the quarter-finals,” says Achim Zeileis, outlining the possible course of the tournament for the favorites. “Spain will probably not face the really strong teams before the semi-finals. Conversely, defending champion England will play in the most difficult Group D with France and the Netherlands.”
Host Switzerland with chances of surprise
The simulations show that host Switzerland has a good chance of making it through the group stage (81.1%). However, the probability of progressing to the next round drops substantially as the tournament proceeds. Norway, Sweden and Denmark are also likely to perform in a similar manner. “With the home advantage and a bit of luck, Switzerland could still spring a surprise,” says Achim Zeileis.
The machine learning model used for the simulations was developed in an international collaboration between several teams led by Andreas Groll (TU Dortmund University), Christophe Ley (University of Luxembourg), Gunther Schauberger (TU Munich) and Achim Zeileis (University of Innsbruck). Besides predictions for sports tournaments, similar ensembles of machine learning and statistical models can also be used in other areas of research, for example for weather forecasts.
Service: The entire forecast with interactive graphics is available here:https://www.zeileis.org/news/weuro2025/