Qatar is the first Arab country in soccer history to host the World Cup. Due to the high temperatures prevailing there, the games will not be played in summer but in winter. From November 20 to December 18, 2022, the world's best teams will be competing for the World Cup title during the Christmas season.
The (statistical) favorite this time is Brazil, with a 15 percent probability of winning. This is what an international team of researchers led by Dr. Gunther Schauberger (Chair of Epidemiology headed by Prof. Dr. Stefanie Klug), Prof. Dr. Andreas Groll and Neele Hormann (both TU Dortmund), Prof. Dr. Christophe Ley (University of Luxembourg), Hans Van Eetvelde (University of Ghent/Belgium) and Prof. Dr. Achim Zeileis (University of Innsbruck/Austria) found out with the help of machine learning. The forecast combines several statistical models for the teams' playing strengths with information about the team structure (such as market value or number of Champions League players) and socioeconomic factors of the country of origin (population and gross domestic product).
Using the values from the researchers' model, the entire World Cup was simulated 100,000 times: match by match, following the tournament draw and all FIFA rules. This results in probabilities for all teams advancing to the individual tournament rounds and ultimately winning the World Cup.
"By repeatedly simulating the entire tournament, you can also see what influences the draw and the tournament tree can have," explains Dr. Schauberger. "It is therefore not always the case that the stronger teams automatically have a better chance of advancing." For example, Germany's comparatively tough group with Spain means it has less chance of reaching the round of 16 than Belgium. After the group stage, however, Germany overtakes Belgium in terms of chances of advancing.
The favorite for the 2022 World Cup is Brazil with a 15 percent probability of winning, followed by Argentina (11.2 percent), the Netherlands (9.7 percent), Germany (9.2 percent) and France (9.1 percent). Despite the fourth place for Germany, the tournament is of course still not over - as shown by the comparatively low win probabilities of some of the top nations. Dr. Schauberger illustrates how great the uncertainty is regarding the outcome of this World Cup: "Our most likely tip Brazil is around 15 percent. Accordingly, the probability that we are wrong with this tip is 85 percent. Only when you add up the probabilities of the top five favourites you end up with a probability of over 50 percent."
So far, the forecasts of the research group have been quite successful: The Innsbruck model of Prof. Dr. Achim Zeileis, which is based on adjusted odds of the betting providers, was already able to correctly predict the EURO final in 2008, as well as the World and European champions Spain in 2010 and 2012, among others. At EURO 2020, which was postponed to 2021 due to COVID-19, France was predicted to win the title, rather than the actual champion Italy.
This year, it will be used for the second time after EURO 2020 as part of a broader combined model developed by the teams led by Prof. Dr. Andreas Groll (TU Dortmund University), Dr. Gunther Schauberger (TU Munich) and Prof. Dr. Christophe Ley (University of Luxembourg), whose predecessor model had outperformed bettors' prediction accuracy at the 2018 World Cup.
The researchers' calculation is based on three different sources of information: A statistical model for the playing strength of each team based on all international matches of the past eight years (Universities of Ghent and Luxembourg), another statistical model for the playing strength of the teams based on the betting odds of 28 international bookmakers (University of Innsbruck) and further information about the teams, for example the market value, as well as their countries of origin, such as population size (TU Dortmund and TU Munich).
These different sources of information were collected not only for the current tournament, but also for previous tournaments up to and including the 2002 World Cup. Using this historical data, a machine learning model is trained that relates the aforementioned information to the respective performance of the teams in the individual tournaments and can thus be used to predict matches in the current World Cup. How well the model performs will be known by the evening of December 18 at the latest, when the final of the World Cup has been played and the new world champion has been decided.
In addition to the scientific interest in the World Cup, the entire team of researchers is also concerned this year with the circumstances under which the World Cup is being hosted in Qatar. "The media reports of the last weeks make clear how problematic the awarding of the World Cup to Qatar is, especially with regard to human rights as well as the working conditions during the construction of the stadiums and the infrastructure," says Prof. Klug. "As a result, a large dark shadow lies over the tournament, making it impossible to follow the tournament with the usual joy."
To the entire forecast with interactive graphics
To the Chair of Epidemiology
Contact:
Prof. Dr. Stefanie Klug
Chair of Epidemiology
Georg-Brauchle-Ring 56
80992 München
phone: 089 289 24950
e-mail: stefanie.klug(at)tum.de
Dr. Gunther Schauberger
Chair of Epidemiology
Georg-Brauchle-Ring 56
80992 München
phone: 089 289 24955
e-mail: gunther.schauberger(at)tum.de
Text: University of Innsbruck/Romy Schwaiger
Photos/Graphics: Pixabay/private/Groll, Hormann, Ley, Van Eetvelde, Schauberger & Zeileis