First of all, thanks for an absolutely marvelous website!
This activity https://intervals.icu/activities/3391545627 The website correctly tells me I spent 97% in HR zone 1, so I know that I’ve been mostly doing around 50% FTP, which would give me a TL of ~45 (which I set manually). But the website calculates a TL of 111, so intensity ~80%, which is clearly impossible at HR zone 1
Perhaps it’s because I don’t have any such low-intensity rides with a powermeter, and the model isn’t built to handle extrapolation gracefully? Because for rides in HR zone 2, the estimates are pretty accurate
Thanks! Your explanation is probably correct. You could change to the HRSS model. That would probably give better results for this kind of ride but “time in HR zones” is likely to be more accurate for rides that better match the ones you do with power.
I see. I guess this feature is not really worth implementing, but maybe you could have some fallback option in the model, where if there is not enough data yet to get a good estimate of the “power in a given zone”, you’d take it from the zone settings. My zone 1 is set to go up to to 55% of FTP, so from there the intensity of the activity I posted shouldn’t be larger than 0.55.
Thats a good idea for the bottom end which is where that model struggles. I have added it to the todo list.