#HACKMCFC, Data Science & Football

After a long, potentially over-stretched UEFA European Championship this summer, football took a well deserved break in July. While gossip around Pogba and Ibrahimovic kept supporters of Manchester United busy, followers of Manchester City took on a different endeavor: #HACKMCFC was the first-ever event, which brought together data scientists and digital enthusiasts from across Europe to answer one question: How can data science help to transform the game?

Surely, the arrival of Pep Guardiola at Man City was one of the major stories this year. After three years at German powerhouse Bayern Munich, the Spanish coach felt the urge to try his talent in a different environment. Guardiola, who is known as highly detail-oriented and data-driven, had implemented his distinguished style of play in Munich with rigor, despite heavily being confronted with criticism along the way by the club’s officials with a rather conservative background.

At Manchester City now, Guardiola seems to tap into an environment much in line with his preferences. In July, the club organized #HACKMCFC – a hackathon, geared towards innovative data products. Building upon in-game tracking, player and health data, participants were challenged to develop smart tools and methods for insight. The winning team convinced the jury with the development of a machine learning algorithm, which tracks in-game decision making. This hackathon is part of a broader data push, as the club has also established a rather large analytics center, which informs training, tactics, coaching and health provision.

Not all football clubs in Europe are that advanced, indeed a lot of the advantages of modern data science techniques are still not being leveraged by clubs and hackathons in this sphere are more of an exception. But with an increasing amount of data available, it might soon be a decisive competitive advantage. Here are the three main areas of application for data science:

Better scouting of talents

In times of intensifying competition and skyrocketing transfer fees, football clubs are increasingly looking to improve their scouting activities. Recruiting top talent at an early age (yeah, that is 10-12 years these days) is essential for a lot of clubs. At later stages, sourcing talent globally – be it in Africa, South East Asia or South America is key. Quickly comparing a player on stats, physical stature and other metrics is essential. But not only comparison is vital – by putting metrics into a developmental context (by having a time series of those metrics, tagged to successful – or not so successful – players) helped club officials to expand their search pattern.

Certainly, judging on a player’s developmental potential purely on metrics is still like a lottery – as many external factors, injuries, squad constellations and the like still have a major impact and are as such rather unpredictable. However, leveraging the power of algorithms to detect patterns of success can help to avoid transfer premium, scout off mainstream and secure a sustainable competitive advantage.

Squad optimization / line up optimization & tactics

With the arrival of big money, fairy tales in soccer are a rare happening. But they happen. Last year’s Premier League title of Leicester is certainly among those fairy tales, heavily beating all the odds. But while the title of FC Midtjylland in the Danish Football League a few years ago might look as a fairy tale on its surface, this time it was everything but luck, as the clubs is known to apply some of the most rigorous data analytics across the entire organisation – on and off the pitch. The clubs owner Matthew Benham, who gained fame with his company Smartodds, is known to bring the data glimpse towards professional sports.

Since Benham took over the club, the team has not only optimized its squad, but also continuously changed its line up depending on the qualities of each opponent. Using sophisticated data analysis, weaknesses of teams can be objectively identified. Is the team especially vulnerable for set-pieces? What habits does the central defender have? While team briefings used to be short and boring, data analytics can create a vital edge for clubs. And the stunning success story of FC Midtjylland in Danish and international soccer only outlines the vast opportunities in this field.

In game decision, analytics, new patterns

Besides game preparation and player scouting, data analytics can inform daily training and coaching routines. By analyzing game data, in-game decisions of players and formation dynamics, insights for daily improvement routines can be detected. This is certainly the easiest, when looking at set pieces. Looking at the high scoring probability, it is quite surprising that set pieces have been belittled by many clubs for quite a while. However, as game video analytics becomes more and more automated, insights provide the stage for significant improvement edge.

But also for more dynamic situations, player behaviour on the pitch can be optimized based on historic data. Coaching players how to move and how to take decisions on the pitch will increasingly be the focus of training. Soccer, more than ever, is a game of avoiding mistakes. Thus, taking the optimal decisions – on a given risk preference – is key for teams, which strive to survive in a competitive environment.

Given the wide range of opportunities of data science and data analytics, it will be exciting to observe, how quickly those techniques disseminate within the sector. Manchester City has set an ambitious example – which teams will follow suit?


Leave a Comment

Your email address will not be published. Required fields are marked *

This site uses Akismet to reduce spam. Learn how your comment data is processed.