Predicting the outcome of the Eurovision Song Contest is something most fans – and many betting punters – try to do year on year. But a computer scientist from Glasgow has turned to a simulation to spot voting trends over the history of the contest, and based on the results, has projected a likely winner for the 2006 edition.
Perhaps unsurprisingly, programmer Derek Gatherer found that there are pockets of biased voting centering around regions such as the Balkans, Scandinavia and the Baltic states. Gatherer claims that these patterns are strong enough to be used as a prediction tool for the outcome of the contest, and makes a bold prediction for the 2006 winner: Bosnia and Herzegovina, represented this year by Hari Mata Hari with the song Lejla.
Simulations of the contest are nothing new. Earlier this year, a team of Austrian Artificial Intelligence researchers put together a program that analysed patterns and gave an identikit-style predicted winner. The outcome of that system suggested the 2006 victor as a Western European man singing a 'happy pop song' – not perhaps how Gatherer's predicted Bosnian winner could be described, except on terms of sex!
The differences between the two simulations are worth pointing out; while the Austrian scientists focussed on many factors, including the style and presentation of the performer, the latest prediction from Glasgow seems based (somewhat cynically!) on voting bias, which does not count for everything.
The true reliability of either system will be put to the test on May 20th, when 24 countries will battle it out for the real grand prix in Athens.
