Each one of the charts that follow is based on the work of well-established data scientists. Coach Mike Norton explains how they might be used in athlete selection, to make a case for changing the athlete’s training protocols, or to define a goal and a path forward.
At best, data analytics help coaches sharpen the focus of an athlete’s training and more accurately predict race performance. At worst, data analytics distract both coach and athlete with benchmarks that may or may not lead to success. Norton calls out functional threshold power (FTP), one of the sport’s favorite metrics, as a potential detractor. Norton points out that while FTP is extremely important, it’s a single reference point. When FTP stagnates, the onus is on the coach to identify new stress that will lead to adaptation. It’s easy to continue training hard, doing the work, but missing the demand that will create results.
Functional reserve capacity (FRC)
This classic chart based on research by Dr. Andy Coggan highlights a significant gap in the athlete’s functional reserve capacity (FRC). As a general practice, I like to review the athlete’s annual plan alongside their training data, and also review data from the athlete’s best racing result.
In downloading the data and seeing how the effort played out, we can tell that the last 90 days of training in the lead-up to the race did not match up with the demands of the race. FRC typically requires a massive number of repetitions at 20–90 second efforts. Because the athlete’s training yielded less than 0.5% in FRC, it’s no surprise that they were unprepared for the 6% demand of race day.
Note that this chart is not showing a measure of FRC/W, but how much time has been spent—or not spent—training within the range that is believed to produce the biggest gains to FRC.
Critical power to interval duration chart
This popular chart, developed by Mark Liversedge based on the research of Dr. Phil Skiba, presents a model that captures telling drops in the athlete’s power output. We can see that the athlete is fantastic up to 7 seconds, then there is a massive drop at 15 seconds. We have to contend with this hole in the curve, which would typically suggest that we do some testing at 15-second intervals to determine whether or not this is truly a limiter. For example, it could be that the athlete did a lot of “standing start” work, which would explain why they were performing so well under 7 seconds; the data could simply be a reflection of the work they’ve put in.
Note that this chart is not significantly different from a standard power-duration curve. It simply offers another way to look at the data with critical power (CP) points along the curve with CP, W’ and Pmax models. This model is more based on Skiba’s work, as opposed to Coggan’s (shown in the previous chart).
Athlete power profile in W/kg radar chart
I like shapes because they relay a story and shapes resonate with people. I can show an athlete a radar chart like this one, developed by Marcen, for communicating a given shape as our goal, and then backtrack to how we are going to get there. In this case, the athlete appears relatively well-rounded, but they are lacking adequate endurance and functional threshold power. The chart clearly tells the story and defines a goal for the athlete.
The connection between human and AI data analytics
There was a time when coaches needed a few weeks to see their athletes and spot the opportunities or gaps in their training. With the right data analytics and a good understanding of the stories they tell, coaches can arrive at the same conclusions or even better ones in a fraction of the time, allowing more time to hone the athlete’s performance.
AI is a tool that can quickly point to possible strengths and limiters, but it only knows what it is fed. This is where the coach needs to use the data to pick a direction to further test and develop the athlete, then keep returning to data. “That’s where sport science is going. We know the answers, we just have to figure out how to get there,” says Norton.