Insights Blog

How to put data and football on the same team

05 Dec 2019 | Share

Our experience with the DFB Akademie (German Football Association)

Must-win game. Fine margins. One-hundredth of a second. These are all phrases we associate with sport at the highest level. They emphasize the importance of winning and how the slightest of advantages can make the biggest difference. These are also topics and outcomes that now rely heavily on data analytics.

Over recent weeks we’ve investigated these issues in-depth with those inside the worlds of football and data, to learn first hand how we can continue to help the two worlds meet. I had the pleasure of participating in the Hackathon set up by the DFB Akademie (German Football Association) surrounding the Germany v Netherlands European Championships qualifier, as well as attending the first Statsbomb conference at the home of Chelsea, Stamford Bridge. From the conversations we’ve had, it’s quite clear that we’re entering a decisive moment in the field of sports data analytics.

Speaking the same language

As part of the DFB Akademie’s ‘Innovation Week’, the DFB Hackathon was a brilliant practical way of combining data experts and specialists from the football industry. There’s so much that both sides can learn. On the one hand, the data scientists and engineers, had the chance to understand the requirements of the football experts in their domain. On the other hand, the coaches and those in the football set up were able to boost their data literacy and awareness of the impact data analytics can have on their job and their passion. As is the case in any industry, ensuring the two sides speak the same language is the key to success. And so it proved in the Hackathon.

Matchday

Finally, after all these years I got the chance to pull on the German shirt and represent my country! The Hackathon itself was divided into seven tasks, with a 2-person team from both Germany and the Netherlands working to provide solutions to different challenges. Each team consisted of a data expert and a professional from the football side. Scouts and match analysts were paired with data scientists and engineers from different industries – this was a real test off effective communication between data-focused teams and the rest of a business.

Challenges ranged from analyzing offensive strategy to set pieces and defensive transitions. We had already been provided with data from a previous Bundesliga game and the Germany v Estonia game, so that we could familiarize ourselves with the task, before the live hackathon.

Then it came to our own challenge, which was focused on set pieces. I’m pleased to say, we won, which contributed to an overall victory for Germany! Compared to our opponents’ pitch, which was centred around creating space at corner kicks, our solution concentrated on assessing the probability of a team committing to zonal or man-marking.

There was so much more to gain from the Hackathon than winning the individual challenges though. It was a brilliant exercise in teaching the data experts to understand the language and requirements of the domain experts, in this case the coaches and scouts, and vice versa. This understanding isn’t a unique requirement for sports analytics. In any business, whether that be retail, healthcare, financial services or any other sector, a mutual understanding between the different teams is essential to getting the most value out of the data at your disposal. And, like so many other industries, the world of sport is at a critical crossroads in its use of data.

Meeting the data inflection point head on

While the hackathon focused on a limited data set concentrating on a specific match, football clubs are now realizing that to deliver richer analytics they have to identify trends over a far longer period of time. Analyzing insights across a player’s 10 year career for example, or several years of fitness data from training, can provide the extra competitive edge that could make the difference in a game. In addition to this, incremental use of positional data is adding more to the load that clubs have to bear.

To support this requires a modern data warehouse, delivering the high performance and scalability to make this a reality. Today, many clubs are working from legacy analytics databases, which cannot handle large data volumes on a live connection. The result is that they will struggle to deliver rich, real-time analytics using this setup. Alternatively, other clubs are relying on using pre-defined statistics from data providers, rather than running their own analytics. Either way, there are challenges in store.

If this technology doesn’t work, data-based insights will be delayed and the coaches, scouts and players will gradually lose trust in the data they are being fed. Football clubs now face a clear choice – to take their data analytics requirements seriously, or lose out to the competition.  The DFB Akademie Hackathon showed the great opportunities for collaboration between the football experts and their data teams, they now need the technology that makes this a reality.

Find out how we can help your team go further with data here.

10 trends impacting data analytics

Now that we’re well and truly in the age of data, what’s coming next? 

FREE WHITE PAPER

Close