TrainerRoad has announced an adaptive training approach with a slick new video:
I’ve been waiting for this kind of development for some time. Platforms like TrainerRoad, the SufferFest, TrainingPeaks etc are all building incredible data sets that know so much about us, it’s a no-brainer that this rich data can be used to better adapt training plans to us as individuals.
With thousands and thousands of completed (or failed) workouts that feature our full data, it should be increasingly possible for this to be used to dynamically adjust our training loads and sessions as our performance and activity levels change over time.
This is an obvious use for “big data” and AI (artificial intelligence).
In my day job, I run Facebook Ads (as well as Twitter and LinkedIn) and in simple terms Facebook has a set of data points on all users that anonimises into a pattern for each of us. In advertising, if I click on a link or ad, Facebook looks via it’s AI/Big Data for other people with a similar pattern and shows them the same ad. This is because my and their profile are similar so if I found a link interesting enough to click on the other person might too.
If we apply the same train of thought to performance training on Trainerroad (or any other platform), the platforms should have a rich enough data set to both use a similar pattern based approach and to adjust loads/workouts when my training performance changes.
My personal health profile: weight, FTP, intensity, training history, HR and power history will create a pattern to work from – either to match against athletes with a similar profile or to adjust as my performance changes.
As an example, I haven’t been riding as much recently so my fitness has dropped and my weight has gone up. Power will be down as well. I did a Zwift ride last night, where I worked harder than normal to low a moderate intensity ride (2W/kg). My heart rate was much higher than normal.
In a world of adaptive workouts (and on a training programme), my next workout should adjust in intensity to offset my drop in performance.
I don’t have a TrainerRoad account active any more so can’t check how well this actually works in practice but this should be the next frontier of platform based training plans.
It’s also a commercial opportunity for those platforms that have deep data like this (I’m looking at your Garmin, Strava and Zwift in particular).
It’s great to see TrainerRoad moving in this direction and I’ll watch to see how it evolves. The only other platform I’ve heard of that does this is a small Italian app called Bike Evo that doesn’t seem to have gained much traction. TrainerRoad huge data set should give it a very strong basis to build from.
Find out more at: https://www.trainerroad.com/adaptive-training/
Thanks for reading.