Big Data and wrong decisions at the EM
The European Football Championship in France once again illustrates extremes. Nevertheless, one might think that game, bread and whip are now subject to fixed scripts, patterns or structures. In the meantime, even scientists are on the verge of predicting the winners among the 24 teams at this year's EURO using Big Data. Will they succeed?
Löw's boys, Busquets' veterans or other hard workers? Spectators won't know which team will ultimately bring home the winner's trophy until the beginning of July. Researchers, however, would like to use Big Data analyses to calculate which tactics are promising and who therefore has the best chance of winning the European Championship.
With the help of Big Data and statistics, scientists at the Karlsruhe Institute of Technology (KIT) are already determining who the most likely European champion will be. Their technology takes new football rules into account. Nevertheless, higher laws also exist. Instead of systems as in the past, tactical "game principles" determine which player should behave how in which situations.
New rules relevant?
"The term game system no longer accurately describes what happens on the pitch today," says sports scientist Dr. Dietmar Blicker from KIT. His research group produces match analyses for Karlsruher SC, among others, and studies tactical developments in football. "Whereas "4-4-2" or "5-3-2" used to be played for 90 minutes, the modern game has become much more flexible."
So the tactical training of players is becoming more and more of a focus. "On the pitch, players must be able to recognise for themselves whether they are needed in the role of a central defender or a full-back. With regard to the new football rules, which will come into force on 1 June and thus also apply to the European Championship, Blicker says: "The changes are not dramatic for the most part, but in individual cases they can perhaps lead to exciting tactical developments in the game."
For example, the kick-off may now be played directly backwards, and an "attacking wall" to obstruct the goalkeeper's view is prohibited.
Germany - no European champion
"The good news from the study is that there is a 94.3 percent probability that the German national football team will get through the group stage and reach the round of 16," says Professor Michael Feindt from KIT. "However, the probability that Germany will become European champion is less than ten percent." France (34.4 percent), Spain (17.5 percent) and England (11.3 percent) have better chances.
Feindt is an expert in big data analysis. The company he founded, Blue Yonder, develops forecasting software and supports retail companies with data-based decisions for inventory planning and dynamic pricing. The prediction software uses self-learning neural networks, recognizes patterns and correlations in large amounts of data and creates forecasts from them.
Cern technology
It was originally developed for use in elementary particle physics at Cern, where it successfully discovers the few scientifically relevant particles among thousands. For the European Championship forecast, the algorithms were adapted and fed with the results of all 36,000 official international matches played since 1873. Of course, many imponderables determine the course of the tournament, such as weather, injuries and wrong decisions.
"Nevertheless, we can use our algorithms to predict the probability of each team leaving the field as the winner in Paris on July 10," said Michael Feindt.
