Evaluating strange forecasts: The curious case of football match scorelines

Abstract

This study analyses point forecasts of exact scoreline outcomes for football matches in the English Premier League. These forecasts were made for distinct competitions and originally judged differently. We compare these with implied probability forecasts using bookmaker odds and a crowd of tipsters, as well as point and probability forecasts generated from a statistical model. From evaluating these sources and types of forecast, using various methods, we argue that regression encompassing is the most appropriate way to compare point and probability forecasts, and find that both these types of forecasts for football match scorelines generally add information to one another.

Evaluating the pinnacle of football match key statistics as in‐play information for determining the match outcome of Europe's foremost leagues

“…Notable contributors to this research include Anders and Rotthoff (2011), Bar‐Eli, Tenenbaum, and Geister (2006), Carmichael and Thomas (2005), Chowdhury (2015), Cifuentes‐Faura (2022), Gasparetto, Mishchenko, and Zaitsev (2023), Gasquez and Royuela (2016), Gómez‐Déniz, Cárdenes, and Sánchez Pérez (2019), Greenberg (2015), Mattera (2023), Mechtel et al. (2011), Reade, Singleton, and Brown (2021), Titman et al. (2015), Turner, Meserve, and Bowen (2001), and Megía‐Cayuela 2023).…”

“…The study findings indicate that the likelihood of winning while playing at home varies across different positions on the pitch, ranging from 41.99 percent for wide midfielders to 91.34 percent for center midfielders. Reade, Singleton, and Brown (2021) compared exact scoreline forecasts in the English Premier League with bookmaker odds, tipster crowds, and statistical models. They found that regression encompassing is the best method for comparing these forecasts, suggesting that both point and probability forecasts for match scorelines add valuable information.…”

A new model for predicting the winner in tennis based on the eigenvector centrality

“…The loss functions used in this work to evaluate the forecasts in this work are the Brier Score (BS, Brier 1950 ) and the log-loss. Further forecast evaluation methods in sports are discussed in Reade et al ( 2021 ). As in Angelini et al ( 2022 ) and Gorgi et al ( 2019 ), we use the Diebold and Mariano ( 1995 ) (DM) test to statistically assess the performance of the proposed model with respect to each competing specification.…”

How rational is gambling?

“…What is more, forecasts based on odds do not seem to perform worse than forecasts based on more elaborated statistical models (see Štrumbelj & Vračar, 2012; Forrest et al., 2005), although there is some work which has found the statistical models to be superior (see e.g. Reade et al., 2020).…”