@Dieter1960 said in Subject: Feedback on Zone Sense accuracy: Short intervals and Zone 2 threshold shifts:
Analyzed the data through a specialized AI model to identify physiological breakpoints
I’d be curious to know which model you used. If you ask an LLM to interpret your R-R data it generally uses the same (or similar) model that DFA-alpha1 is based on, but you can then tweak the prompt to ask it to compute the data using an “adaptive” model that should be more similar to what Suunto uses. At least this is something I tried long time ago, and eventually the computation was too heavy for the model to produce a result. AI models have evolved enormously since then.
Nevertheless, I no longer look at ZoneSense data during my runs, simply because it just seems to “adapt” to my intensity. If I run at my typical aerobic effort, the shift between green and yellow is very likely where I expect it to occur, but if I run something more demanding (trail, tempo, race, etc.) it puts me in the aerobic zone when I actually have ±10 bpm above my aerobic runs.
