Dynamic Bayesian forecasting of English Premier League match results with the Skellam distribution

Robert C. Smit (), Francesco Ravazzolo and Luca Rossini
Additional contact information
Robert C. Smit: Vrije Universiteit Amsterdam, The Netherlands

No BEMPS72, BEMPS - Bozen Economics & Management Paper Series from Faculty of Economics and Management at the Free University of Bozen

Abstract: Due to teams trading players and/or changing their manager and players getting either in or out of injuries and other factors the abilities of teams to score or to prevent an opponent from scoring changes throughout the season in most if not all sports and association football is no exception. As such, we developed a dynamic model based on the Poisson difference distribution, called Skellam, where the scoring abilities are changing over time and are different across teams. The model is developed in a Bayesian framework and is fitted using the Stan modelling language. In this paper, we introduce a unique method used to handle promotion and relegation within the league. The model uses 3 different seasons and the forecasting ability has been measured and validated on the 2018-2019 English Premier League season. As a result, the model predicts the outcome of the matches correctly about 60% of the time. Moreover, we find that the model under-performs somewhat with the best performing, worst performing teams and some of the promotion teams, which could be attributed to both the fact that the season was an outlier in regards to performance for these teams and to the possibility that the hierarchical model may have caused shrinkage.

Keywords: Bayesian hierarchical models; dynamic models; English Premier League; football data; Skellam distribution (search for similar items in EconPapers)
JEL-codes: C11 L83 Z20 (search for similar items in EconPapers)
Pages: [43 pages]
Date: 2020-09
New Economics Papers: this item is included in nep-for, nep-ore and nep-spo
References: View references in EconPapers View complete reference list from CitEc
Citations:

Downloads: (external link)
https://repec.unibz.it/bemps72.pdf (application/pdf)

Related works:
This item may be available elsewhere in EconPapers: Search for items with the same title.

Export reference: BibTeX RIS (EndNote, ProCite, RefMan) HTML/Text

Persistent link: https://EconPapers.repec.org/RePEc:bzn:wpaper:bemps72

Access Statistics for this paper

More papers in BEMPS - Bozen Economics & Management Paper Series from Faculty of Economics and Management at the Free University of Bozen Contact information at EDIRC.
Bibliographic data for series maintained by F. Marta L. Di Lascio () and Alessandro Fedele ().