Dynamic Network 3 -- 0 FIFA Rankings: Replacing an inaccurate, biased, and exploitable ranking system

TL;DR

This paper introduces a dynamic network-based ranking system for international football teams that outperforms FIFA rankings in accuracy, reduces bias, and is less exploitable.

Contribution

The paper presents a novel dynamic network model for ranking football teams, improving predictive accuracy and robustness over traditional FIFA rankings.

Findings

Dynamic network model outperforms FIFA rankings in World Cup prediction

Reduces continental bias in team rankings

Less vulnerable to exploitation compared to FIFA system

Abstract

We explore the advantages of representing international football results as a directed network in order to give each team a rank. Two network-based models --- Static and Dynamic --- are constructed and compared with the FIFA Rankings. The Dynamic Model outperforms the FIFA Rankings in terms of World Cup predictive accuracy, while also removing continental bias and reducing the vulnerability of the FIFA Rankings to exploitation.