According to Rovanos Tsafack Nzanguim, a PhD candidate in Applied Mathematics at the University of Twente and a football fan, the answer is more nuanced. Statistics can reveal a lot, but they cannot replace a crystal ball. “From a statistical perspective, we should be cautious about interpreting any model as a prediction of a single winner,” he says.
One red card can change everything
Football is not a laboratory setting where every condition can be controlled. An injury in the quarter-final, an early red card, a controversial refereeing decision or a penalty shootout can change the course of an entire tournament. That is why Rovanos believes a good statistical model should mainly calculate probabilities, rather than present certainties.
Statistics can help identify favourites. They can show which teams look stronger on paper, which countries are performing well and which squads have the most quality. But they cannot promise who will lift the trophy. Even the strongest team in a tournament does not usually win with certainty.
Football data can tell us more than you think
Klement’s model looks at factors such as GDP per capita, population size, climate and the FIFA ranking. According to Rovanos, these are interesting factors, but modern football data offer many more possibilities. Anyone building a World Cup model today should, in his view, look more closely at the game itself.
He would start with measures of team strength, such as Elo ratings and recent performances over the past few years. He would then add football-specific variables: expected goals for and against, squad depth, market value, age balance, injury risk, the presence of world-class players and the number of players active in the strongest European leagues or at top Champions League clubs. It is not just the quality of the starting eleven that matters, but also who is on the bench and how many expected starters stay fit.
Such a model does not point to one winner, but works with probabilities. Spain would have a certain chance of winning, for example, France another, Argentina another and Portugal another. That may sound less spectacular than “the Netherlands will become world champion”, but scientifically it is more honest.
Three correct predictions are no guarantee
The fact that Klement correctly predicted the previous three world champions makes his model stand out. Still, that does not automatically mean he will be right again. According to Rovanos, statistical models can identify favourites and estimate their chances of winning, but they cannot rule out uncertainty. In a knockout tournament, one moment can make all the difference: an injury, a red card, a penalty shootout or a refereeing decision.
Maybe that is exactly what makes football so attractive. The favourite does not always win. The underdog can surprise everyone. One moment can change a match. And that is why people keep watching, hoping, cheering and debating.
The power of uncertainty
For statisticians, uncertainty is not a problem you simply solve. It is something you try to understand. Models help calculate probabilities and identify favourites, but they cannot remove uncertainty. In football, there is always room for chance.
Maybe that is exactly where science and football meet. Data and statistics help us better understand what is likely. But only once the match begins do we find out what actually happens.
And maybe that is a good thing. Because if the outcome were already certain, there would be no need to play the World Cup at all.
