As I often do, I wanted to give a preview of some research we currently have under peer review. EXACT Sports has one of the largest datasets on athlete development out there so it’s always fun to go into it and do some really interesting research and analysis. We are, after all, a data- and science-oriented organization. Because of our work with over 100 college soccer programs, I decided to conduct a study that would detect patterns of behavioral profiles in the collegiate soccer player. I also used the analysis as an opportunity to re-validate some new algorithms that we’ve integrated into our tools.
The findings were really interesting. I’ll leave the stats talk for the pending paper, but the gist of it was that the analysis uncovered 5 clusters of athlete “types”. The reason why we break into clusters is because different behavioral profiles result in different outcomes. By categorizing athletes into clusters, we can quickly synthesize what strengths and weaknesses the athlete is likely to have based on the cluster he or she is in. Bear in mind that it’s only useful at a cursory level: athletes can defy parts of their cluster (everyone is unique); the clusters only serve as loose templates and guides based on best-fit.
Because our research isn’t published yet, there are some details I’m not able to share. But one aspect of it that I found interesting was our work identifying a cluster of high-striving athletes with low emotional control. I’m sure as a coach you are familiar with athletes just like this. They bring a lot of spark to their competition and expect a lot of themselves, but are at risk of imploding when things don’t go their way. While this may all sound intuitive and obvious (“Of course there are athletes like that!”), what’s so fascinating to me is that we’re able to quantify and classify athletes based on the cluster analysis.