With the final of the European football tournament in sight, in this blog we talk to Jens Melvang, former professional footballer and product manager at Stats Perform – the sports data and analytics company – about the art of data analytics and storytelling in football.
Data has become increasingly important in football over the past two decades, evolving from an experimental luxury at the top of the game to a ubiquitous part of every professional setup. In the early 2000s, data analysis was just becoming established in the best academy setups. Now, it’s an integral part of the game that’s expected for player development and in-game strategy.
At EURO 2020, the craft of sports analytics is being performed as a fine art. The quality and quantity of data, the expertise of those who analyse it, and the insights and stories that unearths are a hugely important for each national team. Few people are as aware of this as Jens Melvang of Stats Perform – after all some of the nations competing in the tournament will be more than a little familiar with the business’ tools.
Football analytics: run the numbers
So, what does this data look like and how is it used? According to Jens, the easiest to quantify is physical data – the stuff that paints a picture of the fitness of individual players.
“Through heart monitors and GPS systems, football associations can track the heart rate of players, the distance that they cover, the number of sprints they attempt, their top speed, and how their playing style affects their fitness levels”, he says.
These measurements give an idea of the general fitness of a player, and gauge how they’d compare against teammates and opposition players. Jens adds: “Quantifying the speed, strength and stamina of players helps managers to place them in positions and systems where they can excel, as well as helping them to avoid ‘the red zone’ – the levels of fatigue that risk injury. As with any organisation, you can set KPIs to hit using this frame of reference.”
Expect the unexpected
Beyond physical data, Stats Perform also help analyse tactical data. This is an attempt to turn performance art into science – to quantify the abilities of players or teams in a footballing context and play in a way that makes the most of their numerical qualities.
Jens says: “One very useful metric used by the top teams is xG: expected goals value. Every shot taken in football has some chance of scoring. xG demonstrates how likely a shot is to be scored if taken by an average striker for that competition. If a player is consistently overperforming their xG – that is to say, scoring more than they’re statistically expected to – it’s likely that they’re very good at shooting.”
Another example is Possession Value, quantifying how likely possession in certain areas is in leading to a goal or a dangerous chance. But football is fluid, and the opposition can change, so how are football staff supposed to read these insights in a meaningful way?
Reading the play
As Jens explains, the value of tactical analytics will always be relative, with some risk involved. “Teams can set up in entirely the right way and, due to bad luck, lose regardless. But on the Loading...data science side, what we can do is collate the highest quality findings possible, identify the values of interest to the coach, and then collate and visualise them in the most compelling way possible. In short, we use the best data to tell the best story.”
Hardware is key to this, with digitisation accelerating year-on-year. Image processing, more sophisticated video camera systems, and more comprehensive physical tracking devices are being introduced to get as accurate a set of results as possible.
Beyond the data collection infrastructure, however, it’s absolutely crucial to be able to visualise and justify areas of interest. Technical analysis staff will receive requests from the footballing staff for certain datasets, and the results need to be presented in a compelling format – ideally as simple and informative as possible. Take the below waffle chart, for example.
Here, we can clearly see how the best creators in the Premier League spark chances on goal – either through corners, crosses, in-direct free kicks or open-play passes.
For the data analyst, therefore, simplicity and clarity are key if the right message is going to be conveyed.
This isn’t just important from a coach’s perspective, trying to understand the meaningful elements of the data to make strategic decisions. It’s also about getting players to buy into the plan. As Gareth Southgate told Forbes in 2019, “when a manager used to ask something, we just did it. Now players want to know why they are doing it.”
Former pros, like Jens, then carry that interest into coaching and analytics ventures of their own.
“StatsPerform is one of those ventures, where we distil a colossal number of metrics and analytics into simple, compelling visualisations,” he says. “Whether it’s collective analysis of a team, a deep-dive into player statistics or retrospective analysis of specific situations in previous games, insightful breakdowns are at the fingertips of an analyst in seconds.”
No matter who you’re supporting or how far your team gets this summer, one thing’s for sure: they’ll be using data to eek as much as they can out of their chances to win. EURO 2020 is set to be the latest milestone in the gradual evolution of data-driven analysis in football, and key to that is an appreciation of the need to make analytics simple and actionable.
Stats Perform is a technology alliance partner of Exasol. The company is involved in sports data collection and predictive analysis, as well as maintaining sports-oriented wire services. Stats Perform has also been increasingly involved in Loading...artificial intelligence and Loading...machine learning. Its clients include media outlets, sports leagues and teams, fantasy sports and sports betting services.