An overview of TrainerRoad’s adaptive learning: The future of the faster

TrainerRoad is a a little extraordinary in the universe of cycling training applications. Lacking the brilliance of gamers in Zwift candies, the unconventional humor and set of riding options that come with Systm, and the personal attitude of a human trainer (which comes at a huge monthly cost) at Training Peaks. But the platform is very effective in fulfilling its sole mission: to make you a faster cyclist.

The platform achieves this through its machine learning tool called Adaptive Training, a system that creates whole-based learning plans that are updated daily with machine intelligence software that meets the unique strengths, weaknesses and limitations of the graphics schedule. rider. The program analyzes each workout by measuring how easily the rider completes each workout.

If, for example, you crush a VO2 max workout, the program will adapt and spit out a harder workout option the next day. Or on a day when driving feels difficult, the program will relax you and provide a little less intense follow-up training. You have the opportunity to accept the adapted program or stick to the initial level of difficulty. The more you use it, the more data it can use to fine-tune your training, something like a Google Nest thermostat that fine-tunes the temperature of your home over time by studying your daily usage patterns. As it tracks you through time, it is sold as a subscription service; you pay $ 20 a month or $ 189 if you buy all year at once.

To get started, TrainerRoad creates a personalized training plan to help you prepare for a future race, ride or event. He asks you, among other things, to choose the type of competition (gravel, mountain, road), the date of the event and your preferred days for indoor and outdoor training. For those who do not have a racing goal, who are just interested in building their fitness, there is also the TrainNow option, in which TrainerRoad allows you to choose from a selection of daily workouts in three categories: climbing, attack and endurance.

Adaptive learning may be smart, but it’s still not smart enough to eliminate the need for ramp tests to determine your basic “functional threshold power” (FTP). This indication of the highest average power that you can maintain for 45 to 60 minutes is measured in watts. These FTP tests are included in the training plan at the beginning of the experiment and then re-tested every four to six weeks to recalibrate the program based on your “progress levels”. These levels of progression are the way the app tracks your growing fitness in each workout area. Determined on a scale of 1 to 10, they are calculated using three methods: machine learning, the already extensive set of anonymous company data collected from millions of completed workouts from other athletes, and your own recent presentation of a workout.

TrainerRoad software can be synchronized with almost any smart simulator or power sensor on your bike.

Photo: Kody Kohlman / TrainerRoad

I liked the adaptive training of TrainerRoad. In my testing, I found it to be efficient, cost-effective, and easy to use. I was also inspired by the podcasts the company produces. I listened to episodes with users, including Masters national champion Jessica Brooks, a busy mother with a full-time, high-level job; US Paracycling Nationals silver medalist Francesco Magizano, who is blind; and David Curtis, a mountain biker who went from his couch to the nine-hour Leadville 100 in nine months.

I tested the app in December in Minnesota after coming out of a four-week cycling break due to minor surgery. Without having a serious training goal in mind, I set up an imaginary 100-mile gravel race for the end of May as my goal. I did my ramp test in the recommended Erg mode; abbreviated to ergometer, this is a mode common in exercise bikes where you allow the exerciser to set the amount of resistance for you based on your pedaling power. During my test, there was a time when pedaling was so easy that I couldn’t turn fast enough to handle base power.

Source link

Leave a Reply

Your email address will not be published.