Football Analytics · Research

Beyond Moneyball:
A Network Approach
to Football

Modern football is won in combinations. Yet the analytics that shape transfer decisions and team selection still treat squads as collections of individuals. Our research programme addresses that gap — extending classical player-valuation models to capture the interaction effects that determine whether a group of talented players actually performs as a team.


The Problem with Moneyball in Football

The "Moneyball" revolution brought rigorous statistical thinking into sport. Player ratings, expected goals, progressive passes — these tools have genuinely improved how clubs evaluate individuals. But they share a fundamental assumption: that the value a player brings to a team is separable from the teammates around them.

In baseball — the sport that inspired Moneyball — this assumption is broadly defensible. In football, it is not. The tactical systems that have defined the past decade of elite football — gegenpressing, positional play, high-line defending — are built on the premise that the right combinations of players unlock emergent capabilities no individual could produce alone.

"The question isn't 'how good is he?' — it's 'how good are they together?'"

Transfer markets, however, continue to price players rather than fits. Clubs routinely invest heavily in world-class talent that underperforms in combination — not because the players are poor, but because the analytical frameworks used to evaluate them had no way to measure relational value.

The Research

Our work extends standard player-contribution models with a network layer that captures whether specific player partnerships produce systematically more — or less — than their individual qualities would predict. The model is validated out-of-sample across multiple top-tier professional competitions.

The central finding is that interaction effects are sparse but real. The vast majority of player pairs show no systematic signal. A meaningful minority do — consistently, and with material implications for performance. Those are the partnerships worth building selection and recruitment strategy around.

Applications

The framework translates into three domains of practical advantage for clubs operating at the highest level:

Application in

Performance

Application in

Investment

Application in

Strategy

The shift in framing is straightforward but consequential. Rather than asking "is this player worth the fee?", the network framework asks "does this player make our squad better — and by how much?" That question has a measurable answer. And it is the question that separates good transfer business from great squad construction.

Research programme — active

This programme is part of the AI+Wellbeing Institute's broader agenda at ICLA — exploring how network-level analytical methods can improve human decision-making in high-stakes environments.