Statistics And Euro2004

Who will win the European Cup in Portugal? Many people have an opinion on this issue, and of course it would be profitable if one could look in a crystal ball and know the result beforehand. Some people expect statisticians to have such a crystal ball, and fortunately, that is not the case. Statisticians cannot predict the result of soccer matches with certainty, but they do have methods to estimate the probabilities of different outcomes (see for example Maher (1982), Clarke and Norman (1995), and Dixon and Coles (1997)). In this note I use the methodology of the Dutch Werkgroep Voetbal & Statistiek which is described detail in Koning, Koolhaas, Renes, and Ridder (2003) to estimate the winning probabilities of the participants of EURO2004. Our model consists of two parts. The first part estimates scoring intensities, that is, the number of goals one country is expected to score against another country. The second part is a simulation/probability model where these estimates of the the strengths of the participants is used to to identify which team is more likely to win the tournament than other teams.