After following a TR plan (masters build for century), ai ftp has dropped my ftp by a handful of watts on two detection days. I do all the hard intervals outside to the letter on an outdoor track. Then once a month during the recovery week, I’ll ride indoors to trigger AI FTP detection. On multiple detection days, my FTP was reduced by a few watts here and there.
what’s the point of doing these plans if my FTP is dropping? I am doing these above FTP workouts and, in my opinion, doing them perfectly adequately (these are like 4 to 9 minute above FTP intervals). Isn’t the point to increase my FTP, not just to complete the workouts?
Am I doing this to myself by allowing FTP detection only during recovery weeks? this doesn’t seem like a fair analysis in my efforts the last few months; it’s definitely a wet blanket on what I thought was a lot of work.
Ideally yes in Build. The Build you did/doing is either not enough volume and/or the workouts are too easy. Particularly if you’re coming off of Base plans that had more volume/TSS. The TR plans, workouts, AI integration, etc. error on the side of caution. I found in my case it was way too conservative and had to use Alternates to chose harder workouts, particularly if I hadn’t done a workout in a particular “energy system”.
The easy answer is to do some kind of test. AIFTPD is a black box. If you don’t feed it relevant data, it’s not going to give you accurate information. As I’ve said here before, AIFTPD isn’t magic or really “new”, as plenty of programs have done this kind of thing over the years. All of them share in common that they’re only as accurate as the inputs (and some are better than others).
Second, the point of all training plans is not always to drive up FTP. In the case of a century, improving your time to exhaustion (how long you can maintain FTP or how long you can maintain an event-relevant power) is probably more important than your actual FTP.
That said, the real issue here is the first part. Trusting any FTP calculator without maximal efforts is always going to lead to inaccurate data. The best thing to do is a long effort at FTP and feel it. Barring that, you can do any number of “long form” tests, 20 minute max effort as an example, and get a better number. Or you can do a ramp test (I’ll leave the editorial comments out on that one. )
Long story short, I did a ramp test today and am finding those results to be in line with the FTP detected number. It says to me, “you’ve completed this base, build, and specialty phase and your FTP is lower, but you did exceed expectations at your event”. So, I will reframe this into being positive for trusting TR’s AI results. Perhaps this is a sign to continue trusting the process?
I think I will keep using the TR plans and AI for day-to-day training for rest of this year, but test with the Kolie Moore protocol; this is what I was doing before returning to TR this season.
I wanted to mention that running AI FTP Detection during recovery weeks does not impact your detected FTP value.
AI FTP Detection looks at all of your recent and historical training data to detect your FTP. In other words, don’t fret – all of that hard work you did is still being “fairly” considered by the algorithm.
That said, it’s good to remind yourself that fitness is more than just FTP. Are you feeling like your time to exhaustion for given power zones is improving? Has your repeatability of hard efforts gotten better? You mentioned that you exceeded expectations at an event recently, which is a good sign that you’re still on the right track!
)