Now that heat training is the last hit or if you leave in a place like me where soon endurance rides will be “hot”, your relatively low power sessions may result in a moderate to high RPE.
I was wondering how TR AI will interpret this. I mean if your ride is at 60% ftp under the scorching heat and your RPE is high. How will the AI react ?
this is rally interesting question, I was wondering the same. There should be someone who have tested it to share experience. I don’t see how it would not negatively impact AI. Probably it will think that your fitness is diving. There should be a place to point that you did a heat training in the survey maybe.
This is a difficult metric to judge. Like said above, a survey question might be the best option. Something to add as a new feature? I’m not sure if TR has any data points to reference Temp from a Garmin will not tell the whole story on outdoor rides. You have to factor in humidity. So that one’s out. Indoors on the trainer can be a factor if it’s located in an area that’s affected by outside conditions. Great question.
I think that there are calculators that factor in temperature and humidity conditions as well as altitude on power. Could a simple solution be that TR AI accounts for it considering your local weather and altitude data, or if you do heat training indoors give you the possibility of manually adding the temperature and humidity of your ride?
It could actually be pretty sophisticated as much as they want it to be. From a development/technology standpoint, it would be very easy to provide weather data minute by minute for every outdoor ride that has geolocation data. There are services/API’s that can give you fairly precise weather (temp, humidity, wind info, cloud cover, precip rates, etc.) if you can give the system a time and location from the past. But that data isn’t free, so I think it more about weighing the benefit vs. cost. A simple optional survey after the workout might provide a similar benefit without paying for all that data.
Whatever the approach, it’s a good idea. I’ve seen the effects of heat first hand this year. We had crazy heat in Feb and March, very few rides where it didn’t hit 80 and many where the garmin was reading 90’s or even over 100f in the sun w/ heat coming off black top. It’s been brutal and I moved all my intervals inside back in March. Most of my endurance work has been outdoors in the heat. The heat definitely cuts into the watts and adds fatigue, but I know it drives some other adaptations. It would be interesting to feed that data into the system and see if there is anything useful there. I just know the heat makes those rides less enjoyable.
I’d been thinking along the same lines, heat has such an impact on performance that it can really derail a ride. However, absolute temperature isn’t a great indicator as we all hold different individual adaptation levels. We have hot summers here, so 40C can feel ok after a month or so … but that would kill me in the middle of winter. Other times, I’ll switch off the fans and do heat adaptation rides on the indoor trainer, the impact of which isn’t related to ambient temperature.
Inclusion in the post-ride survey would surely be best (rather than trying to extract objective data).
The system doesn’t account for anything else, so why would it?
Absolutely a simple question could be implemented yet TR doesn’t even have a profile to set up which would/could adjust for personal differences. Plus, IMO the program doesn’t learn you, it just applies what you did to a million or so other rides and averages you out. No machine learning is applied it appears.
This is a bit of a tricky one.
If you’re talking about just riding when it’s hot outside, that’s somewhat unavoidable, and I wouldn’t worry too much about the seasonal effect of riding in warmer weather. Just focus on doing the workouts when you can and hydrating, fueling, and cooling as best as possible.
If you’re doing dedicated heat training where core temp/HR are your main drivers aside from power, it might be better to log those sessions as unstructured rides rather than TR workouts, especially since you’ll be throttling power as your core temp increases, so following a specific workout’s structure isn’t really practical.
If you find yourself with a struggle survey on a workout that you only struggled with because you were intentionally running hot, you could always use the “I did not struggle” response to let the software know that the workout wasn’t an issue for you.
I know that this is on a list of features that we’d like to build out somewhere, but I don’t know where it sits in terms of priorities.
For the moment, keep true active heat training separate from TR workouts if possible. We don’t have that built into the model yet. 
Let me know if this helps!