COMPSTAT 2018: Start Registration
View Submission - COMPSTAT2018
A0234
Title: Ranking soccer teams on their current strength: A comparison of maximum likelihood approaches Authors:  Hans Van Eetvelde - Ghent University (Belgium) [presenting]
Christophe Ley - Ghent University (Belgium)
Abstract: Different strength-based statistical models are presented which can be used to model soccer match outcomes. The models are of three main types: Bradley-Terry, Independent Poisson and Bivariate Poisson, and their common aspect is that the parameters are estimated via weighted maximum likelihood, the weights being a match importance factor and a time depreciation factor giving less weight to matches that are played a long time ago. Their predictive performance is compared via the Rank Probability Score and the log loss. We will consider two cases: the Premier league as an example of a domestic league and the case of national teams.