Sports organizations have been attempting to find more and better ways to analyze athletic performance for roughly as long as organized sports have been around. Until recently, the coin of the realm had been statistics – the length of time it takes a runner to cover 100 meters, for instance. Eventually, performance evaluators were able to determine related outcome data, such as that runner’s top speed.
The next step was understanding precisely how athletes achieved these times and speeds, considering the building blocks of those performances – example: stride length – in an effort to discover possible correlations. It wasn’t until somewhat recently, however, that the technology existed to accurately measure and compare some of these more granular movements. Although the development of camera tracking, GPS wearables and other legacy technologies allowed for some basic measurement, next-generation technologies (such as Sportlight’s LiDAR) allows for more precise physical measurements and athlete monitoring.
With the application of more science to what previously had been something closer to art, performance evaluators can collect, increasingly comprehensively, front-end data streams (movement analytics) to try to better understand how an athlete produced an outcome reflected by back-end data (statistics). With sports organizations now able to access these technologies while cultivating a growing understanding of how to connect the data to performance, a new question becomes fundamentally important: How much are we able to learn from the combination of movement analytics and statistics?
Taking Statistics to New Levels
Today, the average fantasy sports owner knows that touchdowns scored and points per game – or any other basic outcome data – provide only the most rudimentary understanding of athletic performance. And even as sports statistics have evolved from raw numbers and volume stats to include increasingly smarter percentages (OBP over batting average in baseball) and advanced formulas (PER in basketball), they remain limited in their ability to describe the full narrative of performance, let alone predict future outcomes. The collection of outcome data – even intelligently selective data collection – isn’t the same as performance analysis.
Getting Granular
Breaking down a thing into its component parts to better understand it is hardly a cutting-edge concept, and it certainly applies to more than just sports. Performance evaluators have been examining statistical splits and thinking about high- and low-leverage situations for some time now. They’ve also given a lot of consideration to those aforementioned component skills, such as an athlete’s first step or their stamina over the course of a game or a season. And even many of the more complex questions left to be answered – for instance, what is the correlation between a soccer forward’s change-of-direction quickness and that player’s overall effectiveness? – aren’t exactly new to the field.
What has changed is our ability to accurately measure those more granular component athletics movements. The ability to pinpoint stride length (or shot arc, bat speed and other kinetic movements), combined with advanced algorithms and computing power, is allowing sports organizations to compare data streams – movement analytics and statistics – and begin learning more about how component skills affect broader outcome data.
Determining that your man in goal has the best lateral quickness of any keeper in the Premier League has only so much value. But the ability to measure the lateral quickness of many keepers, tie that data to those players’ outcomes (save percentages) and aggregate them to better learn the precise level of influence lateral quickness has on goalkeeper outcomes is enormously valuable. New technologies such as Sportlight, which will be able to help performance evaluators take more accurate measurements of the front-end data to be married with back-end outcome data, can deliver the qualitative proof of performance – and new levels of predictability – that sports organizations have long sought.