How Goldman Sachs Predicts Spain’s World Cup Victory: Inside the Elo‑Based Monte Carlo Model

Core Model: Elo Rating and Goal Prediction

Goldman Sachs did not disclose its full methodology, but the report mentions an Elo‑based strength metric and a Poisson regression for expected goals, which the analysis reconstructs from the paper and related literature.

Using Elo Ratings to Measure Team Strength

The Elo system, originally created for chess by Arpad Elo, updates ratings based on the difference between expected and actual results, giving larger gains for beating stronger opponents. In the model, Spain leads with an Elo score about 52 points above Argentina and 84 points above France, explaining its top ranking.

Poisson Regression for Goal Forecasts

Goal counts in football follow a Poisson distribution, so the model predicts each team’s expected goals using a log‑linear regression with variables such as Elo difference, recent offensive/defensive averages, attacking talent bonus, team momentum, geographic factors, and historical champion effects.

Elo rating difference

Recent attack/defense averages (last 5–10 matches)

Attacking talent bonus

Team form and morale

Geography (home advantage, altitude, climate)

Historical champion effect

Taking the logarithm of the linear predictor ensures non‑negative goal expectations.

Monte Carlo Simulation of the Whole Tournament

With expected goals for each match, the model samples scores for every game, advances winners according to the tournament bracket, and repeats the entire tournament thousands of times. The frequency of each team winning across simulations yields the reported 26% championship probability for Spain.

Model Conclusions and Limitations

The report acknowledges three main shortcomings: it largely ignores defensive strength, it cannot account for player injuries (e.g., Spain’s star Lamine Yamal’s pre‑tournament injury), and it omits coaching tactics and in‑game adjustments.

Historically, Goldman’s World Cup favorites have missed the mark in every edition from 1998 to 2022. Nonetheless, the model provides a structural comparison framework rather than a deterministic prediction.

Interpreting the Results

As statistician George Box said, “All models are wrong, but some are useful.” The model’s usefulness lies in showing Spain as the most favorable “bet” under its assumptions, not as a guarantee of victory; 74% of simulated outcomes still see another champion.

Understanding the model’s assumptions and blind spots allows readers to gauge its practical value and to treat the 26% figure as a probability map rather than a fate script.