In South Texas (and other areas I’m sure), we’ve been struggling through a very hot summer.
On a low volume plan, I do indoor workouts Monday and Saturday, with an outdoor workout on Wednesday. I feel like AT is having a hard time with me, because my outdoor workouts this summer (with extremely hot weather conditions) are like an instant +2 RPE. A workout that would be a cake-walk indoors on my smart trainer with fans and air conditioning becomes barely doable outside with a “heat index” of 105°F.
I wonder if TR could retrain an existing model or create a second ML model exclusively for outdoor workout suggestions?
Engineers could backfill climate data from web services into all recorded outdoor activities, and use that data as a new ML feature to train a second AT model. Then, for outdoor workout suggestions, they could use climate prediction web-services to suggest a workout based on this new “Outdoor AT Model”.
Outside of the request, using the “Alternate” options and picking a lower Progression Level with the same Difficulty (Achievable, Productive, Stretch, etc.) is one option. You can just use a lower version still closely aimed at the objective, but lower enough to maybe close the RPE gap.
In the same way, you can even consider a bigger drop that would move you down further in the Difficulty (Achievable, Productive, Stretch, etc.) world. Depends a lot on how much you expect the outside to “de-rate” your ability.
Thanks for the suggestion! And I agree, I was leaning towards doing something along those lines for now. Just fall back a bit towards the ‘achievable’ end of the workout spectrum
I think it would be better if on outside workouts he could (either in advance or during a workout) reduce the intensity by 5-10%. I don’t think this effects future workouts in AT. And if it does, when it comes to the survey they could add “heat/altitude” etc option as a reason as to why he found it hard. Thus not affecting future workouts.
