France v Spain. The final as seen by Goldman Sachs and BofA

Every four years, the ritual is repeated: the world’s leading investment banks attempt to predict the economic impact of the World Cup (from advertising spend to tourism expenditure, from new jobs created to GDP growth), whilst not forgetting to leave room for a bit of light-heartedness, with predictions as to which nation will win the trophy.

Typically, this latter process is carried out via the use of models developed in-house to assign probabilities to the results of each individual match, predicting the various match-ups and thus arriving at the final.

For example, Goldman Sachs cross-references data on each team’s historical performance to determine the number of goals scored (also using systems originally designed to rank chess players), then simulates a range of probabilities that a given team will reach a particular round, right through to victory. These are flexible predictions, because after each match day the model is run again using updated performance data to generate new probabilities.

This year, the experience is enhanced by the useartificial intelligence. Bank of America, for example, has ‘fed’ Copilot a lengthy prompt that begins as follows: ‘Act as an elite football analyst and tournament prediction expert. Your task is to predict the most likely winner of the 2026 FIFA World Cup, evaluating all the teams participating in the tournament.” This is followed by specific instructions regarding the evaluation of the teams, the recommendation not to focus solely on reputation but also on form, technical factors and other elements that are usually difficult to quantify, such as team cohesion or the mental strength of the players.

In football, the statistical power of any probabilistic prediction model remains limited, as even the analysts who designed them admit. Will AI be able to do any better at predicting the results of a series of football matches – the very essence of what is human, imperfect and random?