Soccer rankings, based on previous performance of each team, have several important roles with a crucial impact on betting such as to provide an objective indication of the strength of each team (McHale and Davies, 2007). Typically tournaments are scheduled so to avoid the pairing of the best teams from the early stages. They also provide an efficient tool to predict outcomes in order to properly fix the betting odds through an objective criterion. The football betting market is based on fixed odds that generally remain unchanged in relation to bettor demand (Goddard, 2005; Goddard and Asimakopoulos, 2004). The efficiency of the estimates of the bookmakers could add a risk of exposure that may generate ample opportunities to uncover inefficiencies in the market. In the previous literature, the approach of modeling the goals scored and conceded by each team showed to be more flexible than directly modeling win-draw-lose match results (Dobson and Goddard, 2003; Goddard, 2005; Lasek et al., 2013). Bivariate Poisson regression (Dixon and Coles, 1997) is used to estimate ranking models on the Italian Serie A historical data from 1934 to 2016. Promising and encouraging forecasting performance is achieved tuning appropriately the reference period of the data used to estimate the model. Such a model is flexible to the introduction of additional team specific covariates that can improve its predictive capabilities.

RANKING AND FORECASTING SERIE A 1934-2016 SOCCER OUTCOMES THROUGH BIVARIATE POISSON REGRESSION

Roberto Benedetti
;
Alessandro Pandimiglio
;
Marco Spallone
2018-01-01

Abstract

Soccer rankings, based on previous performance of each team, have several important roles with a crucial impact on betting such as to provide an objective indication of the strength of each team (McHale and Davies, 2007). Typically tournaments are scheduled so to avoid the pairing of the best teams from the early stages. They also provide an efficient tool to predict outcomes in order to properly fix the betting odds through an objective criterion. The football betting market is based on fixed odds that generally remain unchanged in relation to bettor demand (Goddard, 2005; Goddard and Asimakopoulos, 2004). The efficiency of the estimates of the bookmakers could add a risk of exposure that may generate ample opportunities to uncover inefficiencies in the market. In the previous literature, the approach of modeling the goals scored and conceded by each team showed to be more flexible than directly modeling win-draw-lose match results (Dobson and Goddard, 2003; Goddard, 2005; Lasek et al., 2013). Bivariate Poisson regression (Dixon and Coles, 1997) is used to estimate ranking models on the Italian Serie A historical data from 1934 to 2016. Promising and encouraging forecasting performance is achieved tuning appropriately the reference period of the data used to estimate the model. Such a model is flexible to the introduction of additional team specific covariates that can improve its predictive capabilities.
File in questo prodotto:
File Dimensione Formato  
RANKING AND FORECASTING SERIE A 1934-2016 SOCCER OUTCOMES.pdf

Solo gestori archivio

Tipologia: PDF editoriale
Dimensione 145.92 kB
Formato Adobe PDF
145.92 kB Adobe PDF   Visualizza/Apri   Richiedi una copia

I documenti in IRIS sono protetti da copyright e tutti i diritti sono riservati, salvo diversa indicazione.

Utilizza questo identificativo per citare o creare un link a questo documento: https://hdl.handle.net/11564/700259
Citazioni
  • ???jsp.display-item.citation.pmc??? ND
  • Scopus ND
  • ???jsp.display-item.citation.isi??? ND
social impact