When Martin Lames reaches for his golf clubs or tennis racket in his free time, there’s no place for analysis sensors or cameras. “At my amateur level, it would be a little over the top,” says the 55-year-old. But once the sports scientist is in his golf laboratory at the Technical University of Munich or analyzing the performance of athletes, he can’t get enough data – be it the decline in running performance of a footballer in hot temperatures, the forces on an ice hockey player’s knee, or the size of the splash when a high-diver enters the water. Lames is a professor of training science and sports IT at TU Munich, and last summer, was appointed President of the International Association of Computer Science in Sport for two years. According to Lames, the increasing use of IT in training and competitive sport will alter the roles of athletes and coaches, and increase the enjoyment of spectators.
SAP.info: Professor Lames, in one of your projects, you analyze official match data from the German professional soccer league. What have been your findings so far?
Martin Lames: As a partner of the German soccer league, we have been responsible for ensuring the quality of positional data for a number of years, which is collected by video. For example, we determine whether the software is able to identify players even in difficult situations such as a close one-on-one challenge, and whether we can be certain that the relevant data belongs to the right person. The data is also used to analyze performance. As a result, we have found that the best team is not the one that runs the most. Take Bayern Munich for example: when their system functions as a complete team, the individuals don’t have to run as much. We attempt to find theoretical models for these observations. Looking ahead to the World Cup in Qatar 2022, we are trying to find correlations between temperature and running performance.
German Bundesliga: Decline in performance in warmer temperatures
How can data from German soccer help you in this respect? We don’t play any matches in that kind of heat.
Even in the games that are played here in temperatures of over 25 degrees, we have found clear indications of a decline in running performance and fewer sprints. Of course, such conditions generally only exist in Germany in May or June, that is, at the end of the soccer season when perhaps some teams have little left to play for and so run less. The challenge for us is to separate the temperature from other influencing factors.
Next page: The new team of computer and coach
This story is part of our special focus on Sports & IT. All the articles related to this topic can be found here.
The fact that we run less or run more slowly in high temperatures isn’t actually particularly surprising.
What we need to determine, however, is when this impacts the attractiveness of the game, or when the temperatures could jeopardize players’ health. I was recently invited to visit the Olympic Academy in Qatar, where I had to the opportunity to speak with soccer officials. They welcome these discussions. One of the viewpoints expressed is that the World Cup could still take place in summer in air-conditioned stadia. They even have large sports complexes where all the environmental conditions can be regulated. I got the impression that money was no issue.
Let’s turn to the real, fundamental opportunities offered by a performance diagnostics system that is increasingly supported by IT. Will talent scouts soon be a thing of the past?
There’s an important differentiation to make. For example, gathering data on running performance is one thing. But observing a team during a match isn’t just about numbers and values. All these help to do is provide a framework. Instead, the analysis focuses on questions like: How willing is someone to go into a one-on-one challenge? How does he use his body? How does he anticipate the challenge? What is his positional play like? We use a qualitative process to find the answers.
Game analyst uses report data to help the coach
While data entry and analysis is improving all the time, more and more of these questions will in future be answered with the help of IT.
It will remain a case of both processes supporting each other. What we would like to see is statistics largely being recorded and entered automatically using sensors. Even the analyses we have today are reports of 60-80 pages. Often, these reports end up on a pile because, on their own, they can’t tell a coach how to solve his team’s problems or how best to use a certain player. However, such reports can be helpful when they are used for quantitative prestructuring. For example, with automatically entered data, you could identify all situations in which the ball was lost again two passes after being won. The game analyst looks at these situations in the video, identifies the causes, presents the results to the coach, and if required, guides the relevant players through a video tactics training session.
One of your projects involves communication within the coaching staff, and an objective is to find ways of improving understanding using IT. Why is this necessary?
One of the trends that we see in professional sport is ever larger teams of specialists working together, including sports psychologists, physiotherapists, nutrition experts, and so on. Increasingly, the role of the head coach is to ensure clear communication, and to establish links between the various items of information that the specialists provide. This is where IT comes into its own. In one of our pilot projects, we are testing our system with next-generation basketball players at FC Bayern.
What does the project involve? What sorts of tools does the head coach need?
Lames: All he needs is simple computer access – and if necessary, a mobile device. This allows him to see at a glance the data entered by the individual specialists. Let’s say a game analyst updates the system with the strengths and weaknesses of individual players from the last match as well as those of the next opponent. This helps the coach decide who to play against a particularly strong header of the ball in the next game. One of the criteria might be the recent jumping performance of his players.
Can’t the coaching staff just talk to each other?
Lames: I’m not saying that they don’t do this too. It’s just that the need for coordination is increasing all the time because more and more specialists are involved and provide their own information.
Data helps athletes understand their own performance
How do the athletes feel about the fact that ever more data about them is being collected, saved, and analyzed?
There are very different opinions. I think that in addition to the new coach profile – working more and more on communication within his coaching staff – we will also see the start of a new player profile. Today, soccer players tend to be very passive: they train as much as the coach says is necessary, and their self-perception is governed by what appears in the media. However, an athlete that understands his performance data will have an advantage in future – not least because he can sit down and study his one-on-one duels that have been analyzed with IT. We want to help athletes “come of age” – an athlete that can study his own data and become autonomous. However, it’s a long learning process.
Next page: Using 3D models to measure stress on joints
How widespread is the use of IT for training and game analysis at sports clubs?
It varies a great deal. Clubs like Bayern Munich already employ four to five people solely for game analysis, and are looking to add more. Others just have a student earning a token wage to do the job.
In which sport is IT used most?
Soccer, without a doubt. Not least because there is already a long history of collecting such data in soccer. Other sports also use technology, but often there is a lack of money. For example, in ice hockey, technological progress is extremely advanced, but only at a handful of the world’s top teams. But then in other sports, they may not even be aware of the possibilities. From a technical point of view, it is perfectly easy to capture a 3D model of an athlete using three simple cameras. You could then superimpose a virtual skeleton on the image, and measure the forces on the knee or other joints. The same technology could be used to measure the body positions of a figure skater, and relay this information to the judges in real-time to tell them who landed best. This would make the awarding of points more objective.
Innovation could help judges by measuring splash in professional diving
Is figure skating open to the idea of such innovation?
I don’t deal very closely with figure skating. But judges at gymnastics tournaments have told me that they find the idea fantastic. Or take the sport of diving, for example: 60% of the points depends on the quality of entry. The judges assess the size of the splash – the primary splash occurs when the diver penetrates the surface of the water, and the secondary splash occurs when the point of entry closes above the diver. In one project, we measured the size of the splashes using video images. Again, the aim would be to assess and measure the formation of the splash in real-time using IT and cameras. You could also identify correlations between the splashes and the angle of entry and turning force.
Next page: USA lead the way with IT in sport
Which countries are most advanced when it comes to the use of IT in sport? And where does Germany stand?
In general, the USA are a long way ahead. Technologies such as image recognition are very well used in sports such as baseball and football. Australia is also at a very advanced stage. Germany on the other hand tends to be very reserved as a whole. However, we’ve also been involved in some projects in goalball – a Paralympic sport – and beach volleyball. But IT is not generally well used for training or competition. In Germany, there tends to be a lack of communication between university research and sports clubs, whereas the situation is much different in Australia where coaches are trained in universities.
A better viewing experience for spectators
Last question: Does the increasing use of IT make also make sport more attractive for spectators?
Definitely. Even today, the live stream is enhanced with figures such as running performance statistics during the game. However, we also see a lot of other information that, for me, is pointless information. For example, we see statements like: Hertha Berlin has not won at Stuttgart for 10 years. But the viewers obviously like this information. With advances in sensor technology, we might see a lot more in future. Sensors in the players’ shirts could tell you about body temperature or tiredness levels. This makes the game more interactive for the spectator. He could form a game strategy: When would I personally make a substitution? The whole sense of immersion – becoming fully involved in the match – would increase significantly. With real-time positional data, for example, you could allow spectators to watch the match entirely from the Bayern Munich player Philipp Lahm’s perspective. That would really put you right at the heart of the game.