AI predicts World Cup glory for Brazil but gives England just 5.7% chance of victory

AI-powered analysis of 38,000 international matches and 1 million tournament simulations gives Three Lions little hope for joy in Russia.

  • Wednesday, 13th June 2018 Posted 6 years ago in by Phil Alsop
An advanced artificial intelligence solution has analysed every international football match played since 1872, running over 1 million simulations of the World Cup and given England just a 5.7% chance of writing their names into football history this summer. Retail AI specialist Blue Yonder has turned the attention of its powerful AI solution to football, and has identified Brazil as runaway favourites for glory in Russia, with a 22.5% chance of winning. Spain lead the chasing pack on 11.1%, with Argentina (9.5%), Germany (9.3%), and France (7.8%) close behind.

 

Christian Haag, football fan and Data Science Consultant at Blue Yonder, explains how the company reached its conclusions: “We spend most of our time using AI to make hundreds of millions of daily calculations to predict customer demand and optimise pricing for multinational retailers, so naturally we wanted to try to predict the unpredictable and forecast the World Cup, using our AI capabilities. We took data from every available international match going to back to 1872 and then ran simulations accounting for different outcomes for each match for all 32 teams, with either Team 1 designated as the winner, Team 2 winning, or a draw. We ran over 1 million simulations and calculated the chances of teams winning or finishing runner-up in the group stage, and reaching each stage of the tournament.”

 

While England fans may not be celebrating in the Luzhniki Stadium on July 15, they can at least look forward to tasting victory in the first match against Tunisia when their campaign kicks off. Blue Yonder’s AI gives them a 65.8% chance of winning against the “Eagles of Carthage”, 20.7% of a draw, and just 13.5% risk of a loss. With a straightforward second match against Panama (71.1% chance of victory), all eyes will be on the clash against Belgium to determine the winner of Group G.

 

Blue Yonder’s AI gives England a 41.7% chance of beating Eden Hazard and co, and a 49.1% chance of progressing from Group G as winners, where they are set to meet Poland or Colombia in the Round of 16. With a 49.7% chance of reaching the quarter-finals, England’s progress starts to look less assured, as the team appears likely to meet one of the tournament heavyweights, Brazil or Germany, and has just a 23% chance of reaching the semi-final.

 

Christian analyses the findings: “One of the interesting predictions that emerged from our simulations was Russia’s stronger-than-expected performance. The team has struggled in recent years, and the lack of a qualification process makes their strength hard to judge, but host teams often fare well, and Russia may surprise many at the tournament. We have calculated that they have a 4.6% chance of winning the tournament, just behind England and ahead of more illustrious names like Belgium and Portugal.”

 

Uwe Weiss, CEO at Blue Yonder, concludes: “Christian and our team have used our AI-based algorithm to analyse a huge amount of data and simulations to formulate these predictions. When not predicting the outcome of football matches, Blue Yonder deploys its AI solution to predict customer demand and optimise pricing for retailers across every product and every store, analysing vast quantities of data to deliver replenishment and pricing decisions on a daily basis. In just 12 months, Blue Yonder optimised replenishment in all of Morrisons’ 491 stores, including 26,000 SKUs across 130 categories. This delivered up to a 30% reduction in shelf gaps and a 2-3 day reduction in stockholding in-store.”