Will computers replace coaches?
Endurance athletes have been asking this question for decades. Why do I need a coach? The easy-to-take answer is that you can’t know everything, and a trained exercise professional knows more than you about cycling. The same goes for your mechanic, who knows just the right wrench to turn to keep your kitten purring. Not anymore.
Pulling your car into the garage, you’re greeted by a professorial-looking fellow in a white lab coat holding a clipboard. He takes your keys and plugs your car into a supercomputer, and out pops what’s ailing it. The days of the grease monkey’s educated guess are over, and we’re better off. Or are they?
Do You Need A TechnoCoach?
Why do we need a coach? The hard-to-take answer is if left to our own devices, we’d get it wrong. Most cyclists don’t know as much about training, when, or how to listen to their bodies. We go too hard when we shouldn’t, not easy enough when we should, and deny it to ourselves.
If only we could plug our bodies into endurance-training supercomputers. An inanimate and emotionless technocoach replaces the value of subjective knowledge and experience. Artificial intelligence-manipulated training guided by wearable tech where algorithms fine-tuned to your body’s responses and adaptations takes the clipboard from your coach’s hands.
Heart Rate Variability-Driven Algorithms Improve Endurance Training Decisions
Why do we need a living, breathing coach? An October 2022 study published in Medicine & Science in Sports & Exercise asks the question differently—your father’s dirt-under-the-fingernails repairman may not like what they discovered. Runners respond more effectively when their training plan is individualized and varied based on heart rate data algorithms than those following a predetermined program.
According to the American College of Sports Medicine, wearable technology was 2022s #1 fitness trend. One key metric provided by wearables, heart rate variability (HRV), is trusted in the endurance community as a reliable predictor of mental stress, physical strain, and overall recovery.
Also Read:
Heart Rate Variability and How It Will Make You a Better Cyclist for a deep dive into HRV!
Detractors of HRV-guided training point to its myopic focus—life circumstances are countless and complicated, and one single data point can’t account for all relevant factors swaying training decisions. The researchers acknowledged this limitation by using three inputs: nocturnal HRV, heart rate response to exercise (the HR-RS index), and subjective evaluation of muscle soreness.
The HR-RS index measures the difference between the actual and hypothetical speed at a specific heart rate. If the athlete runs faster at the same heart rate, the HR-RS index increases and is an indicator of actual monitorable changes in performance.
The Study of Forty Endurance Athletes
Twenty male and female runners were matched into pairs by sex, endurance performance, and training volume. Half of the couples formed the Predetermined Training (PD) group and followed a set plan, and the others became the Individualized Training (IND) group.
All of the athletes began with a similar structure, and the scientists adjusted the IND group’s plan twice a week while the PD group remained unchanged. They started with three weeks of regular training, then six weeks focusing on increasing volume, followed by another six weeks of increased intensity by incorporating one to three interval workouts (6 x 3:00 with 2:00 recovery) per week. The researchers conducted performance tests after each training block, including a 10K time trial.
The study is ambitious, with many moving parts, but the results are intriguing and somewhat unanticipated. First, total training time was almost the same between groups. There was no significant difference in the number of intervals performed between the two, but the IND group showed more variation based on algorithmic plan changes. So the improvements can’t be attributed to which group trained more.
Will Artificial Intelligence take over the role of coaches, or will it be a tool that frees up a coach’s resources to focus on things the machines cannot? Listen to Episode 20 of the Virtual Velo Podcast featuring Coaches Shayne Gaffney and Joy Murphy to discover the future of A.I. for your training!
The Results
Both groups showed improvement, with the IND group performing more than two times better—6.2% to 2.9% in the 10k. The researchers broke it down further to reveal more significant differences in a rather ingenious statistical analysis. The IND group didn’t have any individuals that were non-responders, and 81% reacted so well that the authors characterized them as high-responders.
In contrast, 77% of the PD group had a moderate or worse improvement in 10k times, whereas only 19% of the IND group did.
IND had more high responders (50% vs. 29%) and fewer low responders (0% vs. 21%) compared with PD in the change of maximal treadmill speed and 10-km performance (81% vs. 23% and 13% vs. 23%), respectively.
It is essential to acknowledge that individualized training may allow sufficient recovery and loading to induce desirable adaptations. The study brings out an important point. The objective data and subjective recovery scores prevented non-responsive athletes and increased the high responses.
In other words, accounting for our body’s signals and constantly adjusting based on the feedback seems to account for the unknown life variables in the adaptation equation and take steps toward answering the question. Maybe there is a place for instinctual reaction and value in the experience.
Conclusion—Do You Need A Coach, A Computer, or Both?
Do you need a coach? Yes! Athletes need a view from the outside to get a clear picture. What form that takes may be the question in the future. There is compelling evidence that we are adding input variables to the training algorithm cornucopia. We aren’t close to computers for coaches just yet.
External, objective data has solidified its place and isn’t going anywhere, unemotional and untouched by human flaws. However, athletes aren’t machines that react the same way whenever someone plugs them into the computer. While the study shines a light on future possibilities for the algorithm-driven training pendulum swinging that way, for now, it will settle in the middle between what our coach and we can’t see and what we feel.
What do you think?
Will it come a day when computers replace coaches? Comment below. Your fellow virtual cyclists want to know.
To subscribe to the Zommunique and receive more informative and entertaining articles like this one sent directly to your inbox, click here!
Semi-retired as owner and director of his private Orthopedic Physical Therapy practice after over 20 years, Chris is blessed with the freedom to pursue his passion for virtual cycling and writing. On a continual quest to give back to his bike for all the rewarding experiences and relationships it has provided him, he created a non-profit. Chris is committed to helping others with his bike through its work and the pages of his site.
In the summer of 2022, he rode 3,900 miles from San Francisco to New York to support the charity he founded, TheDIRTDadFund. His “Gain Cave” resides on the North Fork of Long Island, where he lives with his beautiful wife and is proud of his two independent children.
You will read him promoting his passion on the pages of Cycling Weekly, Cycling News, road.cc, Zwift Insider, Endurance.biz, and Bicycling. Chris is co-host of The Virtual Velo Podcast, too!
