Network DNA: which players the build-up actually flows through

Most pass-network charts on football sites look beautiful and tell you almost nothing. Dots and lines. Some thicker than others. You can see Manchester City pass between themselves a lot. So what?

The hidden quantity is centrality. Graph theory has good answers for "which node does the most work routing traffic." Apply that to a pass network and you find out which player the team's build-up actually depends on.

Betweenness centrality, explained briefly

For every pair of nodes in a graph, compute the shortest path. Count how often each node appears on those shortest paths. Normalise. The result: a score from 0 to 1 telling you how much "throughput" runs through each node.

Translate that to football: how often the ball has to pass through this player to get from one part of the team to another.

What it looks like in practice

Manchester City 2024-25: Rodri's Network DNA score is 0.41. That's the highest score we've recorded for any player in any top-5 league in the last 3 seasons. It explains, in a single number, why City's build-up collapses when he's injured.

Real Madrid: Modrić 0.32, Kroos (2022-23 final season) 0.30. These were the team's two pass-conductors. When Kroos retired, Modrić's centrality jumped to 0.38 — partly because he absorbed more routes, partly because Madrid restructured around him.

Where this is most useful

• Spot hidden conductors — players whose central role doesn't show up in goal contributions but is structural to the team
• Predict injury impact — high-centrality players being unavailable hurts disproportionately
• Compare deep-lying playmaker styles — Casemiro vs Rodri vs Tonali: very different shapes despite similar position
• Watch tactical evolution — when a manager changes the system, Network DNA shifts often reveal it earlier than scoreline data

One example of what to watch for

Arsenal's Network DNA shifted dramatically between Arteta's 2022-23 and 2024-25 sides. Granit Xhaka (now Leverkusen) had 0.34 centrality in his final Arsenal season — the highest in the squad. After his departure, Declan Rice took the role and the centrality (0.31 in 2023-24). But by 2024-25, with Ødegaard injured and Havertz fluctuating between roles, centrality fragmented — no player has above 0.22 this season. Arsenal's build-up has become structurally diffuse. Read that into their dropped points and it adds up.

Limits

Pass networks ignore passes that go nowhere. A long ball directly to the striker doesn't traverse the graph in a meaningful way. So direct-style teams will under-show on Network DNA — that doesn't mean they're tactically broken, just that the metric isn't designed for them.

Available on every team's Team Metrics Lab page and on individual player profiles when they meet the minimum 600-minute threshold for the season.