UEFA EURO 2020 Forecast via Nested Zero-Inflated Generalized Poisson Regression

TL;DR

This paper introduces a zero-inflated generalized Poisson regression model incorporating Elo points and match location to forecast UEFA EURO 2020 outcomes, enabling probability-based predictions through Monte Carlo simulations.

Contribution

It presents a novel zero-inflated generalized Poisson regression approach for football tournament forecasting, integrating team skills and match location as covariates.

Findings

Model effectively predicts match outcomes and tournament progression probabilities.

Simulations provide detailed insights into team chances at each stage.

Fitted on historical data since 2014, capturing recent team performance trends.

Abstract

This report is devoted to the forecast of the UEFA EURO 2020, Europe's continental football championship, taking place across Europe in June/July 2021. We present the simulation results for this tournament, where the simulations are based on a zero-inflated generalized Poisson regression model that includes the Elo points of the participating teams and the location of the matches as covariates and incorporates differences of team-specific skills. The proposed model allows predictions in terms of probabilities in order to quantify the chances for each team to reach a certain stage of the tournament. We use Monte Carlo simulations for estimating the outcome of each single match of the tournament, from which we are able to simulate the whole tournament itself. The model is fitted on all football games of the participating teams since 2014 weighted by date and importance.