HOW TO IMPROVE DATA AND DECISION-MAKING USING STATISTICAL ANALYSIS IN FOOTBALL
The tech revolution has modified our behaviour and now it is also revolutionizing sports.
The use of technology in the analysis of sports performance has become a usual and necessary practice among professional teams. Monitoring the athlete’s physical activity allows clubs to know their current physical condition and their playing style, as well as also know the players’ fatigue or the load they are subject to during training sessions. or matches. Thanks to all kind of sensors and cameras we have real-time information of our players position, speed and heart rate data that can also be processed to analyze trends or possible improvements in the game. Two key questions arise at the moment of implanting this kind of innovative solutions: on the one hand, what kind of technology allows best monitoring our players in terms of efficiency; on the other hand, how to process all this data in a useful manner and being efficient with the resources at hand.
Outside the professional world, we are used to seeing all kind of technological solutions created to monitor our advances. Runners can choose between carrying the mobile stuck to the arm to know the distance covered, wearing the latest watch or smartband with a GPS or wearing a tape around the chest with the same purpose. Gymnasts count the repetitions in each machine used or the count of their push-ups and crunches with different apps. However, monitoring elite athletes is where a real technological revolution has taken place. Since FIFA approved in 2015 the use of portable systems for monitoring performance during the matches, the public has been getting used to seeing these characteristic GPS vests under the players’ shirts. Inside and outside the pitch, state-of-the-art technologies are used to know their movements, and the trend is only growing. Different future analyses of the Real Time Location Systems in Sports (RTLS) demonstrate that this is a growing market and it will generate countless business opportunities over the next few years. The comparison is so impressive as easy to understand: a market with a business volume of 280 million dollars in 2018 will turn to 2050 million in 2024. A substantial economic growth that will bring about a more than probable innovation in available technologies.
Currently, the most common sports analysis systems are three: video systems based on multiple semiautomatic cameras (VID), the local positioning systems based on radars (LPS) and the global positioning systems (GPS). The VID is used for the external analysis of sports performance which uses high-definition cameras to monitor players inside the pitch. The interesting aspect of this system is that it allows us to reproduce the tracks all over the pitch, which gives access to some crucial information for the analysis of the movements of the players, not only at an individual level but also regarding the interactions that are produced within the team. Its advantage compared to the two other systems is that it stays outside the pitch, an important variable for many coaches and sports managers who consider that no device should cross the lime line. The LPS systems are above all known by their indoor use, and their operating principle is based on triangulating the position of a person or object thanks to multiple beacons distributed along and across the marked area. However, its use is not as extended as the other two technologies. The use of GPS in sports monitoring has had an exponential growth with the change of the FIFA regulation on the use of technology within the pitch and it is more than likely that it is the most used monitoring system.
Speaking about sports science, VID and GPS are two systems widely used in research associated with the athlete’s sports performance. There are many studies that demonstrate their capacity to improve performance or prevent injuries. In contrast, few studies so far have focused in comparing the use of both methods in official matches as in a recent study in which Eduard Pons, member of the Sports Performance Area at FC Barcelona, has taken part. In this study a total of 759 measurements were recorded. They were associated with the total distance covered, distance covered by minute, average speed and maximum speed during 38 FC Barcelona B official matches in the second division, monitoring twenty-six players of the reserve team. After a statistical and detailed analysis of all of the data, the results demonstrated that, compared to the GPS system, the VID system slightly overestimated variables like the average distance covered or the speed when the players ran more than 6 km/h. An overestimation that, however, did not invalidate the possibility of using both systems and even interchange the information obtained from each of the technologies when analyzing the different variables.
What other factors determine the choice of one of these technologies? Reliability is essential for Eduard Pons, who is also a member of FC Barcelona’s first team. Pons explains that “both technologies are reliable even if it is analyzed together or separately, and this is something that has been confirmed in the paper. At FC Barcelona we have been using both technologies for eight years”. Also, the immediacy with which you can work with the data is a relevant aspect. “Being able to have information from both systems is basic for the load control and the possibility of having real-time information, makes these technologies relevant in the decision making process of coaches”, the trainer says.
System reliability or technology performance is consequently two key factors when choosing what solution to employ in the monitoring of athletes, but there are other variables that, nowadays, seem to be more difficult to measure. The player’s comfort is one of them. Using a GPS vest can be more or less comfortable, but it is something the player feels on his skin. Besides, they can feel parameterized and quantified in every movement they make. Can this sensation affect their global performance? Tomás García Calvo, Professor at the Sports Sciences Faculty at the Extremadura University and Coordinator of the ACAFYDE Research Group commented that “Video Tracking allows us to be able to monitor players without being invasive, and that is an advantage as sometimes elite players are reluctant to play with the GPS device”. The result? “Video tracking technology has allowed us over the past years, to record player profiles and to be able to create conditional profiles in competition in order to be able to use this conditional data later in training sessions where we do use the GPS”. Again, the solution is a combination of both systems.
Another answer to the possible discomfort for the player is the advances in the miniaturization of the devices and the growth of the wearable concept. Roberto López del Campo, Coordinator for the Mediacoach Project & Sports Research Area of LaLiga explains it in a very clear way. “The advances in the so-called smart fabrics and the reduction of the size of these technological devices can eventually evolve to create a simple patch, camouflaged under the badge of the LaLiga logo, and this can eliminate the player’s perception that he is carrying an invasive or uncomfortable device”. Also, he highlights the key effect of the psychological aspect of monitoring in sports: “From LaLiga we consider that football players before athletes are people and as such they function in a completely individual way. There are football players who can perceive a technological device as a threat while for others it is a motivational element. It will depend on each one’s beliefs and perceptions”. What to do in those cases? “The optimal solution is that those who feel better carrying physical devices attached to the body do so. And for the players who don’t feel comfortable for physical or psychological reasons, the optical tracking systems will allow measuring with an acceptable variation the players’ performance without them having a bad sensation for carrying a device”.
In fact, it is also contemplated to monitor the players’ emotions and sensations. Tomás García confirms that “in the same way that we have the quantification of the external load, with the use of technology we can also obtain the internal load where the player is asked for his perception of the effort performed during the competition or training session. Consequently, knowing the elite player’s mood becomes a necessity. Currently, in elite sports, where matches take place every 3 or 4 days, the player’s mental load or mental fatigue is being taken into account as the demand of the game itself affects the player’s cognitive structure. This generates fatigue that is currently being studied because the football player is within an environment with increasing mental stress”.
And what to do with this enormous quantity of data that does not stop growing? The likely decrease in the prices of monitoring technologies, added to the use of artificial intelligence or big data can bring about an enormous revolution in sports. Some imagine the classical scouts being substituted by VID or GPS systems recording all the players information and an artificial intelligence system capable of processing all this big data and identifying patterns of interest or valuable information. All matches could be followed live and with great detail of information if an agreement allowed monitoring the players. Sounds like science fiction? Not for LaLiga, where they are working in this direction. “We have made available for the 42 clubs a virtual environment with all the necessary technology to analyze the (raw data) with this kind of complex analysis techniques, and this new knowledge is being applied on the pitch”, says Roberto. Although he is cautious regarding the applicability in football. “This kind of Machine Learning and Artificial Intelligence techniques (that have resulted positively in other more homogeneous and stable areas of knowledge) have a certain difficulty when applied to football. And we say football and not other sports where the correlation between “good performance statistics” and a positive result is high. The clearest example of this is basketball, a sport in which it is practically impossible to lose a match if you overcome your rival in basketball shots tried, rebounds, recoveries, defensive blocks… Football, however, is one of the few sports in which you can win a match without performing not even one shot to goal (with an own goal)”. However, the final message is a positive one: “We are convinced that the contributions that are currently being made by data scientists, in collaboration with coaches, trainers and football analysts will finally prove the importance that this kind of advanced analytics will have in the prevention of injuries, the rival’s strengths and weaknesses analysis, maximization of the players’ technical-tactical and physical performance, predicting the performance that a player would have if he played integrated in one or another team, selecting the player that would best adapt to a system and determining a playing style…” The future is coming closer.
The Barça Innovation Hub team
Although there are several studies on this topic, many of them have analyzed these demands by looking at just a few variables or using very broad timeframes. A new study completed by physical trainers from F.C. Barcelona has analyzed several of these details more closely.
The understanding of the modifying variables of the game, based on the degrees of freedom.
Sports Analytics has grown exponentially thanks to IT sciences and it also encompasses other subareas (e.g. sports sciences, behavior sciences, medicine or data visualization) in addition to statistics with a focus that is more tactical and sports performance related.