Here is the Mount Lemmon segment. Pretty neat.
I think it would be more accurate if:
- I trained more consistently so it had more useful data in the last 12 weeks which it pulls from.
or - It pulled data from at least 1 year of rider info, ideally more, and weighted it according to how proximal it is to today.
If you do the number 2 option above, you’ll easily have the most powerful predictor tool on the market. I’d be happy to provide input if you’re interested in having a physiology brain comment on the adequacy of your algos.