University of Warwick psychology researchers have been commissioned by the Times newspaper to attempt to predict the Euro 2004 results. Their predictions in “The Fink Tank Predictor” offers grim times ahead for England fans, but the tournament promises to offer many surprises.
The researchers show that England has only a 1.9% chance of winning the whole tournament with a 43.9% chance of not even emerging from group B. However, England fans may gather a little schadenfreude from the predictions that Germany has even less chance than that.
France v. England
According to the statistics the single most likely outcome at the moment is a 1-0 victory to France. Furthermore, adding up the odds across the different full time score predictions shows England only has a 16.4% chance of winning and a 24.0% chance of drawing. That leaves France with a 59.6% chance of coming away victorious from the clash.
Winners and Losers
The chances of each team winning the tournament as predicted both by the researchers and the average bookmaker’s odds say that France is the team favoured to win. The Netherlands is the second favourite. The researchers say the the Dutch have a 24.4% chance of winning, over twice the chance forecast by bookmakers.
Daniel Finkelstein, author of the weekly Fink Tank column in the Times , commented: “Over the last football season The Fink Tank’s forecasts have been a good guide. I am just hoping that this time we’ve got it wrong! The most optimistic thing I can say is that a 1.9 per cent chance is 1.9 per cent better than zero.”
For further details please contact:
Decision Technology Group
University of Warwick
Tel: 020 7289 9174
Press and Media Relations Manager
Communications Office, University House
University of Warwick, Coventry CV4 7AL
024 76 523708
Tel: 020 7782 7485
Tel: 020 7782 7188
NOTES TO EDITORS
The Decision Technology Group (DTG) at the University of Warwick has been developing methods for predicting football matches and tournaments since the World Cup 2002. Their forecasts are the most accurate available based on a track record established through publication in association with The Times’ Fink Tank column over the last two English football seasons. The model was very successful during the last World Cup. The forecasts have beaten tipsters and bookmakers systematically in all three tournaments. The latest forecasts for Euro 2004 are now being made and will be updated regularly over the coming weeks (http://www.timesonline.co.uk/finktank).
The approach to forecasting games and tournaments involves building statistical models of different aspects of the game, such as goal scoring, and then using prior matches to estimate future performance. The model takes account of all qualifying, play-offs and friendly games (around 2,500) for every international team (around 200) to generate its ratings. Each team is rated on two dimensions – “Attack”, their ability to score and “Defence”, their ability to not concede. The model is then used to run a “Monte Carlo” simulation of the entire tournament one million times.
The success of these forecasts represents an interesting case study on the strengths and weaknesses of artificial intelligence compared to human reasoning. On the one hand, the forecasts do not replicate the wide ranging but comparatively approximate reasoning achieved by people. On the other hand, for the narrow range of data analysed by the computer, its calculations have a depth and accuracy that could never be achieved by a human. The model’s strength therefore lies in its ability to impartially integrate thousands of pieces of data simultaneously and then investigate millions of different tournament scenarios.
The Decision Technology Group at the University of Warwick is a spin out company of the university (www.dectech.org) and part of the Institute of Applied Cognitive Science, a state of the art lab based within the university’s Psychology Department. The group was founded in order to provide both bespoke analyses of real-world decision-making problems for commercial clients and more general academic research into human reasoning and behaviour.