Football contains a wealth of data

How good is FC Basel compared to FC Bayern Munich? ETH researchers have developed an online ranking system for top football clubs. This system contains findings on how to improve ranking systems and comprises different approaches to analysing international club football.

Atlético Madrid defeats Olympique Marseille in the Europa League final. ETH researchers’ ranking system enable clubs to be compared systematically. (Photograph: Keystone)
Atlético Madrid defeats Olympique Marseille in the Europa League final. ETH researchers’ ranking system enable clubs to be compared. (Photograph: Keystone)

How surprising was Atlético Madrid’s victory really in yesterday evening’s Europa League final? It wasn’t in the least surprising, considering that the Madrid side is the number six club in the world, whereas Olympique Marseille comes in at 53rd place. Order changed meaning Marseille is still lower than Swiss series champion FC Basel (number 50).

What about the Champions League 2018 final, which will take place in Kiev on 26 May? Which club should be ranked higher, Real Madrid or Liverpool FC? Is the Spanish side really a clear favourite? Or does Liverpool have a good chance of taking the title?

When the Champions League reaches its last rounds or regional championships enter the business end of the season, football fever comes into its own and conversation centres increasingly on the true prospects of the teams involved. For example, how strong is FC Basel compared to Manchester City who knocked them out of the Champions League in the 16th round?

A external pageranking system developed by Professor Ulrik Brandes and postdoctoral researcher David Schoch of the Department of Humanities, Social and Political Sciences, the first system to compare all the teams from the top leagues in the world, may have the answer: Soccerverse, a football universe.

A comprehensive basis developed over decades

Soccerverse doesn’t just compare clubs that have never actually played one another, it also predicts games, includes trainers’ track records and provides a steady flow of new facts on the methodology of system analysis in sport. The site can be used in many different ways and is full of potential discoveries: Bayern Munich is the currently the world’s top club. Of the Champions League finalists, Real Madrid is ranked third and Liverpool FC seventh.

Enlarged view: Netzwerk
The network of all analyzed club teams (Visualisation: Soccerverse)

These comparisons are based on a comprehensive data platform spanning over a million games from 206 countries and just under 3 million goals. A thousand extra games are also incorporated every week. This data pool needs to be managed. At the outset, it was a Herculean task – the system has data going back to 1888 – but now the process is almost self-sustaining: “The data aggregation is fully automated now, it feeds into Soccerverse directly from different internet sites”, says Schoch. This smooth process is due in no small measure to the work of the former University of Konstanz student Imant Daunhawer, who is in charge of the site’s technical implementation and presentation.

That allows the two researchers to focus on further development and the analysis of the results. The results analysis is full of surprises: few would have given Bern Young Boys (currently ranked 80 in the world) a chance of beating FC Basel (ranked 50 in the world) in the Swiss championship at the start of the season.

Strive for improvement

The idea originated in a course held at the University of Konstanz. Potential for improving the current football rankings was recognised and researchers wanted to take a closer look at the processes involved in those rankings. They were particularly interested in the FIFA rankings: “Given the influence of the FIFA rankings on, for example, the World Cup or European Championships, there is a lot of discussion as to whether the ranking process is suitable”, says Brandes.

Enlarged view: Points
The performance development of Real Madrid and FC Liverpool (Visualisation: Soccerverse)

Schoch and Brandes adopted a new approach, transferring the process behind the FIFA country world rankings to club football using amended parameters. “Soccerverse enables us to apply a data-based analysis to establish the consequences of changes to the FIFA ranking system and determine how to evaluate the current system.”

However, rankings are at best a plausible system. They are never entirely objective as they always involve assumptions: “The basis for calculating rankings is invariably man-made. As a result, every process yields subjective decisions”, explains the ETH professor. In spite of this inevitable handicap, it is important to achieve the most balanced solution possible. After all, rankings matter: “The presentation of the results is very persuasive, regardless of how they were actually calculated.”

Material for future projects

In addition to proposing improvements to current ranking systems, the developers are also working to establish new scientific findings. “We basically made Soccerverse as a hobby and we aim to steer it more towards research now. We would also like to offer students the opportunity to work on the project”, says Schoch. Brandes adds that the plan is to incorporate the findings from the system analysis and ranking system in future courses.

The scientists are already optimistic about their chances of prompting improvement measures, even at FIFA: “Our proposals have to be based on real data to ensure their credibility. This prospect is gradually becoming realistic. However, Soccerverse is not a completed project, it is just at the threshold of further development.” The platform thus has a lot of potential, both in terms of research and its influence in practice. As always in science, the process is perpetual: an ever closer approximation of reality.

Comment: The data is valid as at 16 May 2018

Similar topics

JavaScript has been disabled in your browser