the “promise” is if you follow the workouts and rate them how the system wants you to you should achieve the predicted FTP.
I may be getting hung up on the word ‘promise’ but TR doesn’t guarantee that predicted FTP anywhere I’ve seen. Just like a weather forecaster doesn’t for their prediction either.
Here’s what they say about it:
Also, I believe it’s worth clarifying that you honestly rate the workouts how AI thinks you’ll rate them you’ll most likely achieve the predicted FTP (people just putting in the expected rating often leads to strange results). I guess that also doesn’t work for people who have the bug that’s being discussed in this thread.
I agree that it is not a “promise”, I do wonder if some of the issue with people not achieving the predicted FTP is because it is currently weighted towards a best case maximum rather than the most likely FTP. (unless that has changed?)
I think that’s what’s been happening with me. It predicts a 10W gain at the end of the month and I end up with a 5W gain. I’m happy with the progress, but when I compare it to ‘what could have been.’ It’s a little disappointing.
I am glad I found this thread. I was really excited back in the winter when AI FTP prediciton came out. It played nicely with my routine, I was able to aim for something and it worked pretty well. I subscribed for another year in April… And then I hit the outside trails and wow… what a disappointment.
I won’t go into many details but just wanted to show you an example from today. After I finished my ride (it was ok, I was able to hold the power), it dropped 5W which I expected.
The AI prediction was at 281W with 1 workout remaining
Then, I followed one of the advice to maybe replace this one remaining workout with a harder one and I did. What happened with the prediction? It dropped 15W! Based on a single remaining workout that I haven’t done yet!
I am pretty sure that no matter what workout I will do, the FTP will still be around 269, because that’s what I had throughout the summer when I have done only outside workouts.
Welcome to the forum! ![]()
How are your workouts feeling?
I’d be happy to take a look and let you know my thoughts here in this thread if you’re cool talking about your training here.
Let me know either way. ![]()
Yawn. A figure of speech!
Yes I am the same, I am only on TR becuase I have the old gauranteed price.
That is a bit of an assumption (incorrect but common narrative), group and club rides made me the strongest I ever was on a bike.
I make no assumptions…
Research beats anecdote every day of the spacetime continuum!
Happy that you group rides got you strong.
What research ![]()
On the subject of the thread, from my perspective it’s NP that causes a problem.
Yeah it’s pretty hilariously bad.
It feels like you could literally hardcode a few common sense rules to avoid like 75% of the ridiculous scenarios that happen with AI FTP Prediction.
I mean even just as a workaround while you actually fix the thing.
There is close to no scenario where I should do a scheduled workout, rate it exactly what it predicted I would and just see the FTP Prediction get lower over time.
What if the ride data does not reflect the targets or correlate with recorded RPE?
Do you happen to know if the FTP detection button still appears after you’ve set your FTP manually?
I’m not sure how it happened, but at one point the “Detection” button appeared—it showed up in my calendar about a week before the end of the 4-week “Base Season 1” block. I clicked it, and it suggested something like 285W or 283W—pretty sure it was 285W. At the time, I had it set manually to 290W.
I chose not to accept the new value, figuring it didn’t really matter since I was heading into a recovery week anyway. But surprise—after doing some outdoor rides during that recovery week, today it showed my FTP as 280W. Now the detection button has vanished, and it’s nowhere to be found in the calendar anymore.
Up until now, my indoor trainer workouts have been fine with the FTP set manually, but I don’t really want to keep doing that indefinitely. As soon as I head outdoors, discrepancies start popping up. It’s not exactly a minor issue to constantly have that nagging doubt in the back of my mind: “Am I doing this right? Is this too much? What if I end up overtraining without even realizing it?”
Not to mention that, in my experience, my outdoor FTP doesn’t match my indoor (trainer) FTP. Plus, my outdoor power is measured at the pedals, whereas my indoor setup isn’t… There’s no option to maintain two separate FTP values—even though one of the episodes mentions that this difference exists, yet TrainerRoad still doesn’t account for it!
Sounds like a glitch in the matrix.
It will sit there offering you the detected value, and no prediction, until you accept it.
Can you put the pedals on the indoor bike and compare the two? If so, you should be able to see the difference and most likely alter the pedals to match the trainer. I use TPV and it allows you to connect two power sources, I used the scale factor option on my crank to get it to match my kickr, both read within a watt or so of each other now.
there is one. If your hear trate was too high. I am not saying it was, just saying there is an scenario.
If, in fact, your FTP increased from hiking, your next workouts will be easy, and AI wil respond by giving you harder workouts. Problem solved.
Added note: FTP and fitness are related but not the same. If they were, weight lifters, gymnasts, runners, and cyclists would conceivably all have similar FTPs, but we wouldn’t expect that, would we?
What was your criterion for “strong”? Cavendish was strong. Vingegaard is strong. They aren’t the same.



