A group of European statisticians claim to have built a machine learning algorithm capable of predicting this year’s winner of the FIFA World Cup.
The model works by crunching a variety of data on national teams and players before running 100,000 simulations to find the most likely winner.
By this method, the team’s lead author – University of Innsbruck statistician Dr Achim Zeileis – explained in a blog post, the most probable victor of this year’s competition is Spain, with England the next most likely to triumph.
It’s not the first time that the group, comprising researchers from Germany’s TU Dortmund and TU Munich universities, Norway’s Molde University College, and Austria’s University of Innsbruck, has attempted to use machine learning to predict the World Cup winner.
In 2018, the team correctly predicted the US as the winners of the 2019 Women’s World Cup, though they were less lucky with their choices of Spain and Argentina as victors of the 2023 women’s and 2022 men’s competition, respectively.
The algorithm predicting this year’s World Cup works by crunching available data on all national matches over the past eight years.
This is combined with a ‘prospective’ strength estimate of each team, provided by combined odds data sourced from international bookmakers.
The overall strength of each team is then tweaked according to each player’s rating, itself derived from data about their performance at both the club and national level and their expected value in the international transfer market.
This data is finally fed into a ‘random forest’ machine learning algorithm to assess each team’s chances of victory in a variety of different match-ups.
Though confident in the model’s educated guess about this year’s victor, Zeileis warned that it was just that – a prediction, and hostage to statistical shocks like Cape Verde’s recent goalless draw with Spain.
“All our forecasts are probabilistic, clearly below 100%, and hence by no means certain,” wrote Zeileis. “Although we can quantify this uncertainty in terms of probabilities from a multiverse of potential tournaments, it is far from being predetermined which of these potential tournaments we will eventually see during the actual tournament.”
This is not the first time serious attempts have been made to predict the winner of the World Cup. In 2014, the German mathematician Joachim Klement used a statistical model to correctly predict his national team’s victory in that year’s tournament – and then the next three competition winners.
This followed the correct prediction of Spain as the winner of 2010’s competition by an octopus (also German) following its highly accurate match-by-match prediction of the 2008 Euros, a streak that sadly ended with the cephalopod’s death three months after the competition.
Read more:
