As we wrap up this series, let’s recap quickly the milestones on the journey to understand better the nuances of tracking player performance. We will also take a look forward to where this segment of the industry is going and what is needed to get there.
We started by establishing that not all EPTS are created equal and choosing the right one for your club goes well beyond copying what the top teams are doing. It is important to establish the needs of the organisation and what is required to ignite its growth through improving the players’ performance.
In order to achieve this, the first step is to combine condition, workload and tactics, to go beyond raw data and start looking at high order indicators, which “mesh” the work of the various coaches on your staff. This turns the focus towards understanding the connections between the various parameters and moving to insights which can be used immediately to affect a desired outcome.
The key part is getting these high order indicators and insights in real time in the course of a match. In order to make maximum use of them, however, we hinted at a shift in thinking which needs to happen where data is used not just to validate what the eyes see on the pitch, but to actually drive decision-making.
This is where we believe the future will take us – predicting outcomes based on a constant influx of new data. Of course, people wanting to predict the future is nothing new. When it comes to player performance, however, it is new to be able to do it in real time based on “hard” data – data which includes not just the work done by the player but also what cost it exacts from her or him as well as if it (the work done) actually helps the team win.
So, everyone needs better predictive analytics and will continue to have the need for them with much better accuracy than what they are getting now. This is the desired output, the “Holy Grail”, so to speak. However, the input part of this equation – data acquisition – is severely underestimated, and this is just as important as the sought-after output. After all, you cannot have meaningful insights without quality data.
The “fancy” new metric, insight or app feature usually grabs the headlines. After all, as we established already, they can help you understand what you need to improve and lead to a direct result. This has led to most companies in this space focusing on the development of analytics.
Our observation is that while many companies make R&D investments in the performance tracking space, they focus their efforts on the outcomes. They can deploy (or build) machine learning (ML) and artificial intelligence (AI) algorithms which can look for various correlations within the data. However, this makes them dependent on the quality of the collected data. And if you want to have professional analytics, you need professionally collected data – data which is constantly expanding, reliable and robust.
Precious few companies give thought to adding new sensors (e.g. electromyography to track muscle load) or even making their systems into open platforms capable of integrating inputs from multiple sources. In fact, this is what gives context to constantly expanding data. It doesn’t mean simply adding to the mountain of data but adding variety to it. It means adding more data streams, which feed the ML and AI algorithms, thus making them more accurate.
The reason such integration is needed has to do with delivering a true 360-degree service to the team. This means getting the various departments which are responsible for player performance to extract insights from their shared information. As mentioned in an earlier article, this means having a cross-reference among condition, performance, functional and medical data. This includes medical screenings, functional tests, acute and chronic load, leisure activities, recovery and stimulation, and so on.
This makes the outcomes of the algorithms more reliable – increasing the percentage of successful predictions and decreasing the number of false positives. This is an especially severe problem when it comes to injury predictions and fatigue management and control – the main problem we have dedicated ourselves to solving.
Only such a complete approach can give teams stronger understanding and better control over the performance of their players – from fatigue control and management to injury prevention. And we already covered the positive financial ramifications of such improvements both for the club and the players.
But data variety is only one side of the coin. The other one is the quality of the data and this comes from the sensors used to collect it.
Here we underline heavily the importance of using professional-grade sensors. They bring a much higher level of data accuracy especially compared to consumer products. Very often I get the question (even at industry conferences): “How is your device different from the latest smartwatch?” The answer is simple: “We use professional-grade sensors for tracking the performance of professional athletes”, which hearkens back to the earlier statement about having professionally collected data to generate professional analytics.
Shaking your hand or running over 25 km/h while wearing a consumer product like a smartwatch are prime examples. Such actions generate so called “motion artifacts” which create noise and distort the signal your smartwatch receives. Using lower-grade sensors causes the signal-to-noise ratio to drop so much that no algorithm can compensate for it, which in turn affects the reliability and robustness of the data and, by extension, the insights you receive.
This is why we stress the importance of professional-grade sensors in the equipment used by professional athletes where the motion intensity is even greater. This guarantees that the data collected will be the most reliable given the existing technology. After all, we are talking about players who make tens of millions and are worth much more to their clubs. Clubs want analytics to better develop and protect their assets but often take for granted the data used to generate these insights.
Every new generation of athletes pushes further the limits of the human body and rewrites the record books. Soon thereafter, their “discoveries” make their way to the general public (e.g. training regimens, nutrition, recovery techniques) with an ever-expanding desire for data-driven insights. When we look to the future in this area, we see it in smart textiles (and beyond) which integrate a large number of different professional-grade sensors.
This serves to demonstrate that the applicability of this technology goes beyond sports because, besides entertainment, sports serves a much higher social function. Sports is merely the starting place from which we can build towards reaching the entire population in an enormous variety of ways: from social bonding to better health.
To read the preceding parts of this series, click the links below.
Part 1: Not all systems are created equal
Part 2: The importance of connecting condition, workload and tactics
Part 3: The real-time use of insights
Part 4: The value of in-game analysis