Mediaboss Marketing

The way Jordi Cruyff tells the story, it all started with an email. The former Manchester United midfielder was working as the sports director of Maccabi Tel Aviv in 2017 when he was offered a trial with a little-known artificial intelligence company that said it could help the team to prevent injuries. It certainly sounded interesting, and there was nothing to lose, so he decided there was no harm in giving it a go.

The club promptly sent over their squads fitness data and, in return, a warning soon landed in Cruyffs inbox: according to the companys calculations, seven first-team players were at risk of injury. The head coach was not too impressed to be told this, and chose not to act upon it. Five of those players were then injured. It then happened again, and again, and Cruyff soon realised that what was unfolding before him was no coincidence.

Cruyff is now an investor in the company, called Zone7, and more and more clubs across the world are starting to realise that they too might benefit from a technology that could revolutionise the medical side of the sport. With major leagues now preparing for a condensed run of league fixtures, the likes of which they have not seen in the modern era, the need to reduce injuries has become a more pressing problem than ever. And so, as coronavirus put a halt to leagues all over Europe, Zone7 suddenly found they were offering a service that has never been more valuable.

We now have data from over 50 teams across the UK, Italy, France, Germany, Spain and the US, says Tal Brown, Zone7s co-founder and CEO. Brown is based in Silicon Valley and, prior to launching Zone7, founded the first artificial intelligence team at software giant Salesforce. He is not, therefore, what the sport might consider a football person. Neither is the technology what the sport might consider to be football tech.

The instant reaction from many in the game, much like the coach at Maccabi Tel Aviv, is to be sceptical. Silicon Valley technology experts advising football teams how to manage their players? It is not hard to see why there might be some resistance, especially given the stubborn nature of the sport. But the evidence suggests that the Zone7 algorithm works, and that it works spectacularly.

Among the first clubs to sign up were Getafe, who are far from being the most glamorous or wealthy side in La Liga. In two seasons working with Zone7, their injury rate reduced by 70 per cent. Last season, they suffered just eight injuries, by far the lowest in the division, as they punched well above their weight to finish fifth. Atletico Madrid, meanwhile, had 47 injuries. Real Madrid had 32. Javier Vidal, Getafes fitness coach, has described the technology as a tool that will change elite sport.

So how does it work? In its simplest terms, Zone7 uses pattern recognition to rapidly interpret the huge amounts of data that clubs collect on their players. As a sport, we know almost everything that we can know about our athletes, says Brown. We have deployed medical products to measure strength and flexibility, we know how much they are moving around the pitch, we know things about how they play the game. That data is vast.

Digesting it, and converting it into practical information, is therefore a significant challenge. A lot of the interpretation itself is usually left in the hands of the human operator, says Brown. The analysis process relies on these individuals looking at charts, contextualising them and trying to drive decisions.

Using artificial intelligence, Zone7 helps to make those decisions. Rather than a physio having to comb through eight or nine charts per day for 25 first-team players, looking for indicators of fatigue or dips in performance, Zone7s algorithm can do it for them. With all the data stored in its system, it can recognise the patterns that lead to injury.

From there, the clubs are told which players could be at risk and what the next training session should look like for those individuals. An example would be Zone7s algorithm identifying a detrimental pattern in a players sprinting. The club might then be advised that the player should only be doing 400m of high-speed running the next day, rather than 1,000m.

Teams have scientists and physios and strength coaches that are great professionals who know what they are doing, says Eyal Eliakim, the chief technology officer and co-founder. They have to take care of at least 25 players every day in a training programme that is for everyone. It is hard to understand what tweaks are needed and it is really hard to cover all the players. It takes hours and hours every day to manually look at each players data. That is where we come in.

The temptation is to view this as the latest chapter in the data revolution of football, which has led in the past decade to a considerable surge in interest in the statistical analysis of recruitment and tactics. The Moneyball effect, as it is known. On the coaching and medical side of the game, though, things have moved slower until now. The nerds are not taking over, says Brown. But they are making a bigger impact.

The success of Getafe has helped to spread the word about Zone7 and their technology. Leading La Liga clubs are jumping on board and the company is currently in conversation with Premier League teams. Two Championship sides have already signed up, while in the last few weeks alone Zone7 have taken on two new German clients and three teams in Italys Serie A.

Evidently, these sides know the importance of keeping their injury rate low. There is a financial benefit to this, of course, but it is the performance benefit that is so appealing. Getafes success in La Liga tells its own story, although perhaps an even better example would be Leicester Citys title victory in 2016.

For whatever reason (they were not working with Zone7), Claudio Ranieris Leicester enjoyed the most extraordinary season with regard to injuries. Physioroom data shows that in the 2015/16 campaign, Leicesters players lost a total of 275 days to injury, by far the best record in the league. By way of comparison, the league average for the 20 Premier League clubs that year was 1130 days. Being able to consistently deliver high squad availability is a metric that is often overlooked, says Brown.

In their own words, Brown and Eliakim come into this with an outsiders perspective. Their expertise is in artificial intelligence, rather than the sport itself. Indeed, some observers with more of a sporting background remain unconvinced that the platform can have enough information to be truly effective.

I am pretty sceptical of AI and machine learning in injury prediction, says Markus Deutsch, the global CEO of sports technology company Fusion Sport. The AI is not there yet. Its not good enough, and the reason for that is that there is not enough data. You are talking about a very low number of actual injuries over the course of the season. It does not lend itself to good statistical modelling.

Zone7, and their clients, would disagree. And perhaps most excitingly for them, the more teams they sign up, the more advanced the algorithm will be. This becomes inherently better automatically, as it is driven by more data, Brown says. Getafe in season two had better metrics than Getafe in season one, because we had more data. That is also true across the board with more teams.

Inevitably, some of the more traditional clubs will still resist the very concept of distant machine learning as an injury prevention tool. For those who commit as readily as a team like Getafe, though, it could change their approach forever.

Most clubs will just use their data to run their own analytics, which is a great start, says Brown. But if you can have an algorithm looking for injury signals and learning from 10,000 injuries, instead of 100, then that is a different ball-game.

Join our football experts, Jamie Carragher, Jason Burt and Sam Wallace for a discussion and Q&A about the return of the Premier League on Tuesday June 16 at Midday. Sign-up to attend, here.

Read more:
The best way to avoid injuries? Top clubs are turning to artificial intelligence - Telegraph.co.uk