Forecasting of football match results using econometric models


Creative Commons License

Majewski S., Majewska A., GÜNGÖR A. Y.

Journal of Physical Education and Sport, cilt.25, sa.2, ss.338-346, 2025 (Scopus) identifier

  • Yayın Türü: Makale / Tam Makale
  • Cilt numarası: 25 Sayı: 2
  • Basım Tarihi: 2025
  • Doi Numarası: 10.7752/jpes.2025.02038
  • Dergi Adı: Journal of Physical Education and Sport
  • Derginin Tarandığı İndeksler: Scopus, CAB Abstracts, SportDiscus, Veterinary Science Database
  • Sayfa Sayıları: ss.338-346
  • Anahtar Kelimeler: betting market, econometrics, efficient market, Football, match results, travel effect
  • Atatürk Üniversitesi Adresli: Evet

Özet

The issue of foresight in human activity plays a non-trivial role, particularly when financial considerations are involved. The ability to anticipate future events can provide an investor with a sense of security regarding the investment or market advantage over other market participants. There appears to be no such model that can be successfully used for forecasting and planning. Using the terminology of financial markets, the bookmaking market would be classified as an efficient market if such a model existed. Inefficiency leads to inequality and therefore an informational advantage for some market participants over others, which should result in the possibility to realize supernormal gains. Even from the perspective of market organizers, inefficiency that induces investors to gamble or increases the bookmaker's revenue seems more advantageous. The primary aim of this study is to test the feasibility of building a robust econometric model against which predictions of match results can be made. Thus, an indirect verification of the intermediate efficient market hypothesis for the sports betting market will be carried out. The study will be conducted concurrently for two autonomous football leagues: the Polish Ekstraklasa and the Türkiye 1 Super Futbol League during the years of 2021 and 2022. A series of panel regression models were utilized to test the hypothesis. The results of the analyses showed that it was possible to estimate the probability of winning (bookmakers’ expectations).