I would like to use AI FTP, however I do all of my workouts using TrainingPeaks Virtual and match them afterwards. AI FTP requires 10 workouts using the TR workout player. After that all is good and it will continue to update afterwards.
My question is, what is the fastest way to get AI FTP auto-calculating? Can I queue up a bunch of small workouts? Can I load workouts up but skip the intervals?
I was in the same about a month ago when I signed up for TR after taking about 6 years off from it. Those short workouts donāt seem to count or factor into anything. So, yeah, donāt waste your time with those short workouts. I tried doing 5 five minute ones and TR acted like they were non existent.
I started taking my 90 minute workouts and splitting them into 30 minute workouts. That seemed to work, though TR didnāt count the last 30 minute workout but did take my last 60-minute SS session.
It sounds like completing any TrainerRoad workout that isnāt an FTP assessment or āfreerideā of sorts should count towards the 10 workout requirements.
@gloscherrybomb maybe those few sub-5 minute workouts are classified differently which is why they arenāt counting. You might have to step it up to the 15-minuters.
Question for behind the hood on the 10 rides bit. If you just feed the 10 rides with ābogusishā rides like the intro one 10 times, how would that ever generate a reliable ftp? If its using additional rides as well for the calculation then why the heck do you need 10 TR rides?
If I were you, Id just create 10 custom workouts then match them to past rides.
Im not subbed so I cant test it but I think a shortcut to this could be to create the workout in trainingpeaks scheduled for tomorrow. Sync TR/TP then I think you can find that workout in the library for you to match against a past ride.
Matching doesnāt work. All my workouts are TrainerRoad workouts, to the letter, matched to the prescribed workout once uploaded. These donāt count.
Iāll try a 15 minute one in the player, by the wink above it sounds like that may well be the threshold.
I donāt know official reasoning but I am quite sure AI FTP needs N (=10) good/real workouts to classify athlete into some internal profile bucket so it can start estimating how athlete progresses with any kind of future work that probably doesnāt need to be TR workout after that.
Theorizing further, Iād say TR has manually reviewed/classified existing library workouts (PL, plus whatever internal indicators) to complement AI FTP model. And we know automatic PL assignment for imported workouts are buggy in some cases (one known combination to inflate PL is adding high power sprint at the end of interval). This is possible reason why those workouts are not taken into account (yet).
I think it would be good to have clarity on what the requirements are for a workout to count towards the 10, as the support documentation is not correct.
Secondly it would be a good if either the ā10ā requirement was removed completely, or it was changed to also include matched workouts (which is a feature suggestion, so probably not within your remit. You could pass it on though).