Are All Football Leagues Equally Predictable? – Introducing Two New Metrics

Rui Martins, University of Lisbon

Abstract: In sport analytics, particularly in football, evaluating the predictability of outcomes across leagues is essential to understanding competitiveness. Traditional metrics often focus on the likelihood of specific results, such as home-wins, draws, or away-wins. However, analyzing the probabilistic structure underlying these outcomes, particularly concerning the strongest team, has the potential to offer other insights into leagues predictability. This study builds on previous work by Lopez et al. (2018), which introduced the idea of a league parity measure reflecting the dominance of stronger teams. While insightful, this metric primarily captures the predicted probability of winning, rather than also addressing the variability of the game outcomes. To overcome this limitation, we propose two new  metrics: League Outcomes Predictability and League Probabilistic Dominance.