I’m relatively new to TrainerRoad — started about 3 months ago while preparing for a gran fondo I raced a few weeks ago.
When I first imported my data, my detected FTP (227W) was very close to what I had estimated myself (around 232W). The plan felt well structured (3 phases), and fairly quickly the new AI started giving me higher FTP predictions — sometimes quite optimistic, though they usually got adjusted downward over time. I was consistently completing all workouts at 100%.
After the race, I took a recovery week and then left another week open before starting a new plan. During that free week, I decided to do an indoor workout and picked a hard one since I was feeling motivated — 15x1-minute high-intensity intervals suggested by the AI. About halfway through it got really tough and I had to lower the intensity to finish it.
From that single workout, the AI shifted from predicting an FTP increase to predicting a decrease. That abrupt change surprised me a bit.
Eventually, AI FTP detection gave me 251W, which felt too high, so I did my own test and got something closer to 240W. Since then, even though I haven’t failed workouts, the prediction trend has been drifting down (now around ~245W).
Overall, the predictions and detections feel a bit erratic to me.
In the end, I think FTP prediction can create some unnecessary anxiety if taken too seriously. I’m starting to treat it as just another data point and focus more on consistently executing the plan.
The important part is that my fitness does seem to be improving gradually — so I’ll keep going and see how it evolves.
That’s basically all the forum was talking about for the last months: folks taking the prediction too serious instead of acknowledging that it is just a possibility for them to reach this value and not even a very likely one (it’s the very best case scenario IIRC). They’ll probably improve this eventually in some way (Monte-Carlo-Simulations were already mentioned somewhere).
In that sense, your approach is perfectly fine: Stick with your plan, do your workouts and improve over time. The process for success hasn’t really changed with the new predictions. It can be a motivating little gimmick or even a utility to learn more about where your training may take you and how to train better (thinking about why the AI thinks you’ll worsen your progress can be quite insightful)
Thanks for sharing. It sounds like there is likely some stuff I can touch on here based on what you’ve said already. I’d be happy to dig into your career a bit and let you know what I see.
Let me know if I can share that stuff here. Otherwise, we can start a discussion elsewhere.
I’ve just taken a quick look, and I can see that you’re putting your threshold workouts effectively at the end of your week (Thursdays, but Fridays are your only day off). I wonder if you’re coming into those sessions with some fatigue.
Also, it looks like you skip at least one of your hard workouts each week in favor of endurance or unstructured riding, which will definitely have an effect on your progress.
I’m also seeing that your second hard day is sometimes getting adapted down to an endurance ride due to fatigue.
It’s really important to prioritize those hard workouts since they are major drivers of your adaptations and improvements. Is there any room to move things around in order to ensure you’ll be performing at your best for those?
Another thing to be aware of here is that we do have a current bug that affects FTP Predictions for athletes who do some/all of their workouts outside. Essentially, it gives some inflated FTP predictions early on that slowly settle down as they near their FTP Detection. Since you have your threshold workouts scheduled as outside rides, and you often do a majority of your weekly riding outside, you’re likely seeing the effects of this bug in your predictions.
All in all, you’ve made some good progress over the past couple of months, and many athletes start to have to work harder for the same gains once they get to your w/kg. It’s certainly possible to keep moving forward, but I think addressing some of the points I’ve made above will help you moving forward.
Hi Eddie, thanks for the detailed feedback — really helpful.
Regarding the structure of my previous plan, it was probably a mix of a suboptimal setup (I was still learning how to configure and adapt things) and a bit of eagerness to ride more — it was summer here and the weather was great. Most weeks I actually felt pretty good, maybe even too good, and didn’t feel much fatigue.
One of the biggest changes TrainerRoad has brought for me is actually doing less, allowing proper recovery, and still making progress.
My new plan now has 2 hard days, so I’ll make sure to manage fatigue better and keep an eye on any adaptations that might replace those key workouts.
The FTP prediction bug you mentioned sounds exactly like what I’ve been experiencing. The predictions were very optimistic, but even the detections with the new AI felt a bit exaggerated in my case. I started at 227W (imported), then 229W a month later, but after switching to the new AI it jumped to 240W, and then 251W — which feels too high. I tried holding ~105% for 20 minutes and it was clearly not sustainable.
Now the prediction is trending toward ~245W, which seems more realistic — so we’ll see how it settles.
In any case, the important thing is that I’m clearly seeing progress, so I’ll keep focusing on consistency.
