Player Load Evaluation in Sports (PDF)

Welcome back. In this lesson we're going to discuss more about the use of external measures and how they're used to evaluate training and performance. In our last lesson, we were describing the separate measures of accelerometry and GPS, speed and distance. We described how catapult uses the sum of...

Welcome back. In this lesson we're going to discuss more about the use of external measures and how they're used to evaluate training and performance. In our last lesson, we were describing the separate measures of accelerometry and GPS, speed and distance. We described how catapult uses the sum of the acceleration data to revive the measure of player load. In this figure from a recent scientific article, we see a comparison of the player load and GPS distance. The top two plots, show us the relationship during competition. So actually, during gameplay. In A we're seeing the total and in B we're seeing the player load per minute. In the bottom two plots, we see a similar relationship exists during training. We can see that these two measures are correlated, which should make some sense if the activities of the sport relate to the distance traveled. In the sport of soccer with an expansive field, the two are highly correlated. This result is quite a bit cleaner than the last slide. A main reason I expect, is that this dataset provides data from a single athlete focused on running. In soccer, there's considerable cutting that introduces changes in acceleration that are not contributing directly toward the velocity of the athlete. However, in this data set we see a tight correlation between the velocity determined by GPS and the player load determined by catapult device, using tri-axial accelerometry. Note that the line that we see here, we have a line of best fit which looks pretty strong. But what you'll note here is a very clear demarcation between walking speeds here and miles per hour, two to four miles per hour. We have a big transition when this athlete begins to run. In addition, it's worth noting that over here we have some unusual values towards this plot, and the issue is an important one. So, with acceleration, if we are jumping up and down vertically and having no distance traveled, then we are zero in regard to the miles per hour. And yet we might have considerable player load according to catapult. But what we see is that these values are effectively similar to a walking activity. One more thing to note is this sort of curve, a linear response here, where the fastest running that this athlete was doing, falls considerably off that line, similar to the way that walking does. Now that we have an appreciation for what player load is describing, we want to go a level deeper to think about ways which player load might be expressed. For example, to say that someone had a high player load tells us part of the story, but not the whole story. It provides a depiction of the acceleration load of the workout or game, but it does not give us full information regarding the intensity of the workout or game. We see on this slide, the use of two terms. The acceleration load, which is defined as the accumulation of absolute acceleration values. And the acceleration density, defined as the average of absolute acceleration values. Often the measure of average player load per minute is used to help capture the density of the player load measure. For example, you might argue that an athletes top speed in a game is an interesting performance metric that she could evaluate as part of a big data search, for factors that determine top speed. However, there are subtle and possibly less subtle issues that can influence these intended performance measures. On this slide, we highlight to such issues. First, weather conditions can have a heavy influence on measures like peak velocity. For example, snowy soccer or football fields are unlikely to yield the top speeds of an athlete. Similarly, the type of surface, for example, grass or artificial turf, could also have subtle influence on the speeds attained. Second, the style of play of the opponent may lead to misinterpretation regarding low player loads or slower velocities, due to the pace of the game. Similarly, the score differential can play a role in the speed of the game as well. For example, if a team slows down the speed of the game when they have a lead. Next, we want to address the possible differences between practices and games. For example, do you anticipate that gameplay is higher intensity than is performed in practice, or as practice intended to reach the intensity of gameplay? This likely depends most on the particular sport and the duration intensity of gameplay. Let's consider some examples. A long practice may approach or surpass the volume of game, but not the intensity, for example, to reduce injury in a practice setting. However, there is typically a goal to provide game like intensities within a practice as well. Here's some real data collected from a team that had a relatively short and lower volume practice, as well as a higher volume and intense practice. And we compare it to game data from two 45 minute halves. Note, that the total volume of the higher volume intense exercise is greater than the game, but it is also an hour longer in duration. The player load per minute during the game time is higher than either practice, but the max velocity is about the same. These devices, which can collect data from each and every player at every practice and every game, have become tools for coaches and trainers to use to help them prevent injury or at least to reduce the number of injuries. This will be the focus of our next lesson.