Extended modelling of the result of football matches

by Tatjana Slavova

Abstract

In this paper we present the advantages of the modelling of the soccer game based on the additional statistics data on shots on target. We have implemented Poisson models of Karlis D. and Ntzoufras I. published 1998 to evaluate the prediction of number of goals in a match. The parameters of the model include the difference between the offence strength of one team and the defence strength of adversary team in dependence whether the play took place at home or away. The evaluated values of strengths of a team on basis of goals may be used for goals prediction and thus for the calculation of the probability for each team to obtain some fixed position in the final table.In 2004 Beyko, Grozea-Helmenstein, Helmenstein and Slavova have proved the efficiency of the Poisson models for 27 world-wide upper soccer leagues. Meanwhile we have developed this model for evaluation of the strengths parameters of the team using statistic data not only of goals but also of shots on target. Parameters of the new developed model are more suitable for forecasting of the match results as the parameter defined in log-linear Poisson model that accounts only goals. In this paper we present quantitative measure for the amelioration of forecasting results for the upper soccer leagues of Germany, England, Spain, Italy and France.

Unless otherwise stated, the content of this page is licensed under Creative Commons Attribution-ShareAlike 3.0 License