Machine Learning Algorithm Forecasts Spain as 2026 World Cup Favorite, US at 1% Win Probability

A machine learning algorithm has provided its predictions for the 2026 Men's World Cup, following its successful forecast of the 2019 Women's World Cup winner. This modern data science approach aims to offer probabilistic forecasts for the tournament's outcome.

The algorithm operates through a two-step process. First, sophisticated statistical models are combined with expert insights from bookmakers and transfer markets to assess the strengths of all participating teams and their players. In the second step, a machine learning algorithm integrates these strength estimates with other team-specific information. This generates a probabilistic forecast for each potential match, simulating match outcomes like a pair of 'loaded dice' where different probabilities are assigned to the number of goals for each team. For instance, in a simulated opening match, Mexico had an average of 1.9 goals, compared to South Africa's 0.7, leading to Mexico having a 65% win probability.

To determine the tournament's most likely course, the algorithm simulated the results of each match 100,000 times, considering the official tournament draw and all FIFA rules, including overtime and penalty shootouts. The results position Spain as the leading favorite with a 14.5% chance of winning the title. England and France follow closely, each with a 12.4% probability, while Germany holds an 11.2% chance. Portugal and Argentina are also identified as strong contenders, with probabilities of 8.9% and 8.2% respectively. The expanded 2026 tournament features 48 teams and five rounds in the knockout stage, contributing to a tightly packed group of favorites.

The host nation, the United States, has a strong chance of advancing to the Round of 32, with the algorithm assigning a 78% probability, the highest in their group. However, their chances significantly decrease in the knockout stage. The probability for the U.S. team to secure a home victory in the final, scheduled for July 19 at MetLife Stadium in New Jersey, is predicted at 1%.

The machine learning algorithm's simulations are powered by a comprehensive dataset. This includes retrospective estimates of team strengths based on national matches over the past eight years, prospective strength estimates from international bookmakers' odds, and player ratings derived from their goal contributions at club and national levels. Additionally, the current quality and future potential of players, reflected in their expected market values from Transfermarkt, are incorporated. Other relevant inputs include team-specific details such as FIFA rank and the number of players in the Champions League semifinals, alongside country-specific socioeconomic factors like GDP per capita. A random forest machine learning algorithm is then employed to determine the relevance of these features for actual World Cup results.

According to Fortune, these predictions offer a data-driven outlook on the upcoming global football event.