Delta and Delft Integraal/Outlook often write about innovative ideas that offer big promises for the future. But what has happened to such ideas a couple years on? What for instance has happened to dr Rob Kooij’s soccer model?
Delft Outlook, December 2008
Kooij intends to discuss possible applications of this model with the Dutch Soccer Association: “Visualising the networks of other national teams wouldn’t be too difficult either.”
Holland will probably beat Japan next Saturday (June 19) during the World Cup soccer tournament. And this is not just wishful thinking. The match winner was determined by a mathematical model developed by dr. Almerima Jakamovic, dr. Rob Kooij, and one of their students, Fatemeh Badinrad, of the faculty of Electrical Engineering, Mathematics & Computer Science. This model is a refined version of a model the researchers developed two years ago, just prior to the 2008 Uefa European Football Championship.
“The players on the Dutch team have played together very often”, says Kooij, who believes the Dutch players are therefore much more in-tune with one another than the Japanese players. Based on facts like these, combined with the world ranking points Fifa accords to each national team, the model predicts that Holland will win the match.
In order to make this model, Kooij, who also works at TNO, had some students enter the line-ups of all the Dutch international teams in a matrix, which accounted for a total of 700 players, starting with the players who played in Holland’s first-ever international match, held in 1905 and played against Belgium.
Kooij recently applied for a 20,000 euro grant from the Union of European Football Associations, in order to also develop matrices for other countries. This should greatly improve his model.
Kooij has also developed a simple application that soccer fans can use to scroll through all the data he and his colleagues have gathered about Dutch players. This application is accessible on the TNO website (Kooij had previously contacted the Dutch Soccer Association (KNVB), in order to also make this application accessible on the KNVB’s website, but without success). He wants to add many more data mining options to the online application. “During what minute does a certain player score on average? Data like these are also nice to add”, Kooij says.
By clicking on a particular player, the user can also see a line up of all the other players that a particular player has played with and what the results were; for example, Holland’s dream team of Van der Vaart, Sneijder, Robben and Van Persie have won eight out of the eleven matches that they played together, while drawing twice and losing once in the other three matches.
Such information is perhaps also interesting to know for gamblers involved in World Cup betting pools. Who for example will score Holland’s first goal? “Chances are it will be Van Persie, as he scores 0.53 times on average every 90 minutes,” Kooij says. “Huntelaar has a higher scoring rate – 0.74 – but he’s a substitute player.”
www.oranje.tno.nl/oranje-statistieken
Het onderzoeksbureau Labyrinth deed in opdracht van de website dedecaan.net onderzoek naar de studiekeuze van scholieren. Ruim tienduizend leerlingen uit de bovenbouw van het vmbo, havo en vwo vulden een vragenlijst in.
Meer dan de helft daarvan bleek ontevreden met de begeleiding op school: ze moesten veel zelf uitzoeken en kregen weinig persoonlijk advies. Maar de scholieren gaan zelf ook niet vrijuit. Zeventig procent weet dat ze hulp kan vragen aan een decaan, maar slechts een klein deel doet dat ook.
Havo- en vwo-scholieren hebben relatief meer moeite met het vinden van een opleiding en zijn ook minder zeker van hun uiteindelijke keuze: één op vijf blijft twijfelen. Ze vinden het aanbod te groot, of weten nog niet goed wat ze willen.
De resultaten zijn geen verrassing voor Emiel Willms, voorzitter van de scholierenbond Laks. “Dit onderzoek bevestigt nogmaals de noodzaak van betere studiekeuzebegeleiding”, zegt Willms. “Het is schandalig dat leerlingen nog steeds de dupe worden van slechte begeleiding, terwijl de problemen allang bekend zijn. Bovendien is het zonde van het geld.”

Delft Outlook, December 2008
Kooij intends to discuss possible applications of this model with the Dutch Soccer Association: “Visualising the networks of other national teams wouldn’t be too difficult either.”
Holland will probably beat Japan next Saturday (June 19) during the World Cup soccer tournament. And this is not just wishful thinking. The match winner was determined by a mathematical model developed by dr. Almerima Jakamovic, dr. Rob Kooij, and one of their students, Fatemeh Badinrad, of the faculty of Electrical Engineering, Mathematics & Computer Science. This model is a refined version of a model the researchers developed two years ago, just prior to the 2008 Uefa European Football Championship.
“The players on the Dutch team have played together very often”, says Kooij, who believes the Dutch players are therefore much more in-tune with one another than the Japanese players. Based on facts like these, combined with the world ranking points Fifa accords to each national team, the model predicts that Holland will win the match.
In order to make this model, Kooij, who also works at TNO, had some students enter the line-ups of all the Dutch international teams in a matrix, which accounted for a total of 700 players, starting with the players who played in Holland’s first-ever international match, held in 1905 and played against Belgium.
Kooij recently applied for a 20,000 euro grant from the Union of European Football Associations, in order to also develop matrices for other countries. This should greatly improve his model.
Kooij has also developed a simple application that soccer fans can use to scroll through all the data he and his colleagues have gathered about Dutch players. This application is accessible on the TNO website (Kooij had previously contacted the Dutch Soccer Association (KNVB), in order to also make this application accessible on the KNVB’s website, but without success). He wants to add many more data mining options to the online application. “During what minute does a certain player score on average? Data like these are also nice to add”, Kooij says.
By clicking on a particular player, the user can also see a line up of all the other players that a particular player has played with and what the results were; for example, Holland’s dream team of Van der Vaart, Sneijder, Robben and Van Persie have won eight out of the eleven matches that they played together, while drawing twice and losing once in the other three matches.
Such information is perhaps also interesting to know for gamblers involved in World Cup betting pools. Who for example will score Holland’s first goal? “Chances are it will be Van Persie, as he scores 0.53 times on average every 90 minutes,” Kooij says. “Huntelaar has a higher scoring rate – 0.74 – but he’s a substitute player.”

Comments are closed.