Background: Male, 54, recreational MTB rider in decent shape. Just completed a 4-week MTB – General Fitness plan (3x cycling/week), with an initial plan to also do 2x/week short full-body strength training.
What happened:
At the start, the AI predicted an FTP gain of +11W (214 → 225W). Two things unfolded:
Strength training felt like it was hampering my cycling, so I dropped it after roughly the first week. The strength sessions were uploaded from Garmin, so TR could see that load.
The AI FTP prediction kept declining week over week, eventually reaching 0% — and my AI FTP detection after completing the plan confirmed I stayed at 214W.
All cycling sessions were completed as prescribed. RPE surveys were filled in honestly throughout: Easy for endurance rides, Hard/very hard for the VO2 Max sessions and anaerobic On-Offs. No skipped workouts, no intensity reductions.
Questions:
Why did the predicted gain erode to zero despite consistent completion and appropriate RPE ratings? Could the early strength training fatigue — even though I dropped it after week 1 — have suppressed any FTP gain?
I’d like to reintroduce strength training going forward, but I’m wary of it interfering with cycling gains again. What’s the best way to integrate 2x/week full-body strength into a 3x/week cycling plan to minimise interference? Should I prioritise same-day sessions?
Happy to share more data if useful — would really appreciate any insights from people who’ve navigated this or from TR staff.
Did you notice any of the workouts changing for easier variations, or do you think the plan stayed the same and you completed the workouts that were on the schedule to give you the 11 watts.
My guess is the AI felt you needed a slower ramp rate and changed some of the workouts for easier variations.
It must’ve seen that you weren’t completing or responding to the workouts, so it’s thinking you’re not improving for some reason. Are there any activities that you are not sharing with TR? You seem to have missed a vo2 session on the first Monday, and the anaerobic that drops from 4 to 2.5 is the biggest indicator things aren’t going well.
The basic rule with workouts is for them not to effect your next planned workout, so if those are upper body as indicated, I’d consider doing them after your anaerobic and vo2 sessions rather than the day before.
Hers a few more questions;
What’s your longer history with TR?
Did you use the Plan Builder?
Are you sleeping and eating well?
Are you fuelling the vo2/anaerobic work?
Is anything else wearing to down?
As I recall it actually did change some of workouts for easier variations, so you are probably right. However, I would have expected to see at least some small improvements after 4 weeks of adherence.
I am sharing all activities with TR - I actually did the ‘missing vo2’ the day before, so that shouldn’t matter much. Concerning the drop in the anarobic you are probably right.
Good idea - I will try that.
I just started TR after having been of it for a couple of years, however all my training is synced from Garmin, and all are with both power and HR, so TR should have plenty of data on me
I did use plan builder
Sleeping and eating well I would say.
Yes I am bringing fuel for those
Not really - I just got a cold sore though, so maybe I have had a little virus in the body.
Irrespective of all of these I would have expected to see at least a small improvement over 4 weeks of adherence though.
You’re following a similar trajectory to me (and lots of others)
You’ve probably rated a few of your workouts Very Hard when TR was expecting you to rate them as Hard. As a result, it’s adjusting your workouts (and as a result your power targets) to get your upcoming training to the most effective level. A consequence of this is that your prediction drops.
It’s disappointing (until you rationalise it away) but you’ve (we’ve) done good work and it’s setting you (us) up for progress over the longer term, which is way more important than just 4 weeks in Jan/Feb.
not gonna comment on the ai ftp portion of it, but just as an overall plan, I’m not sure I would expect 3.5hrs volume with a vo2 workout and anaerobic (anaerobic capacity is its own thing and not really used for increasing FTP performance) to really move the needle for increasing FTP all that much
You probably did have improvements in your high end power. But not FTP as you weren’t training for improvements there. This is where a predicted 5 min power to go with the planned 20 and 60 would be nice.
in the past FTP wasn’t adjusted during the specialty period. So that could explain the 0%, but not sure why it would show to start.
Do you feel you’ve gotten fitter/stronger over the last 4 weeks? TR’s ftp is just an estimate and is primarily used to help normalize a productive workout range. Yes, it can be a metric to help measure fitness as well, but it’s far from perfect for that.
My original ftp under the new model was artificially high compared to a traditional physiological ftp definition. TR was predicting a small increase, but it ended up unchanged (but is still higher than my definition of ftp). But I have gotten significantly fitter/stronger over the last 4 weeks and that’s what matters.
FTP is not equal to fitness, particularly this new flavor of ftp that TR is using. It’s easy to get caught in the trap of gamifying your training around ftp, it’s such a common metric that people throw around. But if you really want to measure changes in performance, it’s worth finding other metrics to do that. As great as many of the new features are in TR, I personally think the new TR ftp calculation has taken a step back in the context of being a fitness measuring stick.
You are probably right, and I didn’t choose a plan focusing on FTP, but rather on general fitness - but I was nonen the less surpised to see no improvement at all. But then again I probably should’nt put to much emphasis on the estimated FTP.
I find it hard to evaluate my progress on the hometrainer, and the weather in Denmark has been so crappy lately I havent ridden outside for a while now. Hopefully I will experience impovements when I can get outside again. I can appreciate your point on FTP, it is just always nice to see improvements in some metric, and for lack of a better measure I tend to look at FTP, or VO2 from my Garmin - but that is probably not much bettter (that didn’t improve either).
This thread is gold. So much good advice and a user who’s willing to listen and learn. I just want to add another comment to say your general fitness has almost certainly improved from all that VO2 work. Like others have said, that might not show up in the AIFTP just yet, but you are absolutely healthier and stronger than you were before the plan. Stay consistent and positive. You’re killing it!
I will definitely try and stay consistent - I have started a new plan, aiming at improving my fitness for an upcomming MTB trip.
I will say that it appeared to me that the ‘prescribed’ workouts actually were perfect effort-wise. I have tried other platforms, where some workouts were way to hard for me, but it feels like the AI has a pretty good idea of my abilities (or lack thereof).
Knowing you are going to get a hard workout on hard days and an easy workout on easy days, really helps with consistency. The next piece would it learning what type of hard workouts do we prefer.
Mine did something similar. First day of the plan start was 344W to 351W. Over the course of the plan, the prediction kept ticking down and down, even as I was completing workouts.
Pessimistic view? TR gives a nice big FTP gain to make you feel better and keep training even though it knows it will just adjust down over the course of the plan. So it’s not that you’re doing anything wrong, there was just no FTP gain to be had. Meaning that TR should have shown a 0W gain from the start, but for marketing purposes it gave you a bump to make you feel good and sell the product.
According to TR, workouts should never feel Very Hard. If you rate a workout Very Hard it’s going to mess with all your upcoming workouts.
100% agree with this. Especially a large jump that was predicted. The real problem is that AI predicted any FTP gain at all. And in my mind, it’s like ChatGPT where it just wants you feel all warm and fuzzy and won’t do anything to make you upset or disappointed. It’s pandering.
Regardless of the FTP number, I take comfort in the fact that the sessions felt appopriate. FTP is just one metric, and a 4-week block at 54 may simply not be enough to move that needle, especially with the added early fatigue from strength training. And I guess if some of the sessions were very hard it should ramp down.
I’ll give TR the benefit of the doubt for now. The initial prediction was made before the strength training fatigue muddied the signal, so perhaps there was a legitimate basis for it at the time. Whether the model is accurate enough to predict gains reliably is a fair question though. I am excited to see what the results of my next block will be.
