The Barcelona football team of the 2009-10 season were widely regarded as the best the game had ever seen.  The team won the Spanish league and the UEFA Champions League among a haul of six trophies captured that season, all whilst playing a hugely attractive brand of football.  It’s an achievement unmatched in the history of the game, and understandably it attracted envious glances from those wishing to emulate them.

Despite the appeal of ‘Moneyball’ style analytics in the game in recent years, it’s perhaps fair to say that network science was not used by many to try and explain the team’s success.  That’s precisely what new research from the Universidad Rey Juan Carlos in Spain attempts to do however.

“We combine the use of different network metrics to extract the particular signature of the F.C. Barcelona coached by Guardiola, which has been considered one of the best teams along football history,” they say.

Networks on the pitch

Each player on the pitch is represented as a node, with passes then used to connect each node.  As the number of passes increases, so too does the strength of the link between nodes.  The researchers captured data not only on the passes made, but also where on the pitch they were made.

The researchers attempted to use networks to understand the way the game transitions at various times.  They begin by crafting a network from the first 50 passes of the game, before monitoring how this network evolves as the game progresses.

They did this for all 380 games played by the 20 teams that compete for the Spanish La Liga championship during the 2009-10 season, and calculated the network features for each team.  One of these features is the clustering coefficient, which observes how triplets of players pass to each other.  This rating was significantly higher for Barcelona than any other team in the league.

The average shortest path throughout the network was also much shorter for Barcelona than any other team, while the eigenvalue of the connectivity matrix, which measures the strength of the network, was also much higher.

By measuring how the network evolves during the game, the team believe they can identify the network metrics that improve the probability of a team scoring a goal.

Productive patterns

For instance, the data suggests that the way a teams average position on the pitch changes over time is crucial, as Barcelona played much higher up the pitch than most other teams.  Their extreme possession-based game also resulted in them making far more horizontal passes than any other team as they probed for openings to attack.

Despite the evident strengths of the team, the analysis also highlighted a number of weaknesses.  For instance, when the team were dispersed from Xavi, their midfield hub, they were more likely to concede a goal, which is a tactic that other teams eventually exploited to counter-attack effectively against them.

Given that the hegemony of the team was eventually surpassed suggests that teams and opposition coaches were able to identify a way of overcoming the Barcelona style of play without such analyses, and of course the analysis overlooks the fact that the Barcelona team had several players who are widely regarded as among the best of all time, including Lionel Messi, who many regard as the best player ever.

As players such as Carlos Puyol, Andres Iniesta and Xavi have retired, the team has been less successful, which underlines the importance of the right players as well as the right tactics.  Nonetheless, as data and analytics becomes an ever greater part of sport, this kind of analysis is likely to become more widespread.