Last month marked Quanter’s second birthday, and the beginning of a third year of the daily collection through a mobile app of subjective data from players in professional and youth sports teams. This player monitoring system’s core variables relate to training load (sRPE) and recovery factors, as well as injuries & illnesses. While not yet quite on the ‘Big Data’ scale of Google or Facebook, the growing dataset is certainly large, consistent, and our research into the prediction of injury/illness risk is becoming more and more compelling. But before we could get to Quanter’s predictive analytics, the validity of the system’s methods needed to be examined. We are very pleased to say that we have now performed this research; and the results dictate that Quanter’s simple subjective methods must be considered an important tool in the monitoring toolbox, either independent of, or better yet alongside, the quasi-objective measures of HR monitors and GPS trackers.
Before explaining this validation research, a brief reminder of these simple methods. Quanter employs decades of sports science research on players’ subjective self-reporting of workload: the sRPE method (Session Rate of Perceived Exertion). sRPE describes the overall mental and physical fatigue experienced by a player/team. The sRPE value is created by asking players on a 0-10 scale what they perceive to be their overall sensation of fatigue from a training session or match and then multiplying that rating by the duration of the session. These kinds of methods, often using a pen, paper, and one of too many spreadsheets, have been in practice for years across the elite sports industry, from English Premier League clubs to National Olympic Committees. With a mobile app like Quanter, the whole data collection and analysis process is automated, giving coaches the information they need to make decisions immediately. Simple as that.
The RPE component of Quanter’s ‘Training Load Query’
So, how have we validated these methods in Quanter? Working closely with one of our clients, Vaasan Palloseura (a professional football team here in the Finnish Veikkausliiga), we measured the loads of training and matches over a season with both Quanter and the industry-standard HR & GPS sensors of Polar. After normalising the data for each individual player, we performed a correlation analysis on the data consisting 1901 observations. The results, shown below, are compelling: sRPE as measured by Quanter has a large correlation (using Cohen’s criteria) with all the major HR & GPS variables. So Quanter’s subjective data correlates with both internal load measures and external load measures, such as distance, accelerations, and decelerations. Most notably, Quanter’s sRPE has a very high correlation of 0.80 with Polar’s Training Impulse variable (TRIMP, which describes the overall load based on heart rate). And with heart rate measurements often lacking the sensitivity to capture effectively the load of short sharp bursts, the fact that sRPE correlates with external load measures is particularly interesting.
Correlations between Quanter’s sRPE and Polar HR & TRIMP variables
So Quanter’s sRPE method responds to differences in workload in much the same way as these HR & GPS measures: if Polar would say that a training session loaded your player heavily, so will Quanter. While our marketing team might want to describe this correlation as unprecedented, extraordinary, or revolutionary, it is not, and there is no real surprise here; decades of research have already pointed to the validity of sRPE methods in general. The goal in our research was to validate specifically Quanter’s sRPE measurement against quasi-objective measures, a goal that has been achieved. Our research shows how the Quanter app can provide similar overall analyses of workload as HR/GPS systems, but in a more accessible, easier to interpret way, and at a fraction of the financial and time cost.
Correlations between Quanter’s sRPE and Polar GPS variables
Having shown these correlations, and so established the validity of Quanter’s subjective methods in terms of their association with certain ‘quasi-objective’ measures, we then performed a factor analysis on the same dataset to better understand the relationship between the data produced by these different systems. We found that the holistic measurements of the Quanter app can partly explain differences in player workloads that wearable devices, which isolate factors of performance, cannot detect. Of course HR & GPS measures also provide detail not captured by subjective methods, and we are not at all claiming that Quanter should replace these measures. Rather, we recommend coaching teams employ a multi-dimensional approach. Further, we propose that the complexity of quasi-objective measures means that the daily collection and analysis of subjective data, especially through the now validated Quanter methods, is at minimum the first and easiest step for teams of all levels and ages looking to monitor their players’ workloads and recovery effectively.
All in all these results are an excellent start to our research. We have strong evidence for the association of Quanter’s methods and quasi-objective measures. We have strong evidence of the importance of taking a holistic approach to capture the range of relevant information. More importantly, we have a strong dataset from which to model the risk of injuries and illnesses. These predictive analytics, which may well become a game-changer for coaches and players at all levels, will be the subject of our next Quanter Research blog post.