Agree. This is my issue with the way it works. The system has all of my data. It knows exactly how my body(HR) reacts to my watts over years of rides. Why should it depend on me answering a survey to determine if the ride was easy or hard? Especially when we are told beforehand what it SHOULD feel like to us. I feel a lot of pressure to just respond Easy when maybe it was moderate because I don’t want my training to get stalled out for example. Don’t leave it up to me and my human frailties if the system is really so smart that it can give me the right workout every time.
My response was in context to the OP in which he said the workouts are spot on for him.
As far as your wife goes and her RPE in response to the workouts, it’s a bit confusing. If she or anyone rates a workout as Hard, yet she is capable of much more, then why would it be Hard? I don’t get that. Has she referenced this guide:
Maybe she would be better served if she answered the survey in terms of: “Could do the workout again easily, could do another set of intervals, could only do one more interval, could not do one more interval, or Barely made it to the end” \
I sometimes think back to my college days when I trained with The Cyclist’s Training Bible for a few years. I was completely self‑coached—heart‑rate based workouts, RPE, and a handwritten training journal. Funny enough, those were some of my best training and racing years, using tools that feel like the dark ages compared to what we have now.
RPE isn’t a measurement of fitness; it’s just a gauge of how hard an effort feels. But it consistently correlates with your fitness level, and when you learn to trust it, it becomes a surprisingly reliable guide.
We are currently working on a fix for this.
Where did you hear that we weren’t going to fix these issues?
Did you get that figure before or after your race?
The AI version was released in January. It’s now July, and there has been substantial criticism and backlash from a significant number of users.
TrainerRoad has made some welcome tweaks, including restoring some of the control that was originally removed. However, much of the communication over the past seven months has appeared to focus on explaining or reframing what “AI FTP” means, rather than addressing the underlying concerns many users have consistently raised.
Seven months is a long time in the software world. From a user’s perspective, it’s reasonable to conclude that this is largely the direction TrainerRoad has chosen and that the core design is here to stay.
I’d genuinely like to be wrong. Please prove me wrong.
I’ve been a subscriber since the very beginning, and I want TrainerRoad to succeed. That’s exactly why I care enough to be critical when I think it’s heading in the wrong direction.
That “How to rate your effort” guide will be extremely helpful. Thanks for that.
I know it sounds confusing but I can understand her difficulty with RPE sometimes myself when trying to determine whether a workout was hard or moderate.
I’ve been training long enough to have a good feel for it. Easy enough to confirm with some intervals right around threshold as well as over/under intervals. And I don’t get too hung up on my exact FTP, it just needs to be close enough so that threshold intervals are actually threshold, O/U’s are truely O/U’s, and Sweet spot is really sweet spot. All that stuff is self-correcting, you are going to know it pretty quick if you are trying to do long threshold intervals 10 watts over your physiological threshold. And I’m not saying FTP is completely useless as a performance metric, I just find it less of a performance proxy compared to looking at power curves for longer durations.
If you’re referring to the issues with FTP Predictions and Detections, we are currently working on a fix.
While TR AI did launch back in late January, it took time for this issue to surface. The models we use are complex and take time to work with. We really are doing our best to get this resolved as quickly as possible.
We listen to our athletes and take their criticism seriously, and it often plays a role in decisions that are made behind the scenes.
We use FTP as a training benchmark. That likely won’t change anytime soon. The issues that are being discussed here are a result of a bug in the software, not how we classify FTP.
The morning before the race. The night before when I went to sleep was still the number shown for days. The week before the race was a tapering week, so not challenging and when I wake up on race day it was 5W less. I just ignored it and kept to my race strategy as planned and was able to keep the wattage on long climbs (it was a XC marathon).
That is a perfectly valid point. The second sentence of his post however indicated that he was “well trained before” and “wanted to see if (he could) push (himself) further” so in the absence of other information I made an assumption.
Fair enough though, I take your criticism.
Some people don’t like discomfort. Any amount of pushing themselves feel “hard” even if it’s only sweetspot and below threshold. It should be a 7 out 10 and pretty doable for anyone.
I’ve never been able to push myself hard enough in a vo2 effort to puke or even feel like puking. Others can get there. Does it mean that they can just push harder than me? I don’t know the answer.
Not everyone feels ill after a sufficiently hard effort. Different types of efforts will vary in how you experience failure.
Most fatigued I’ve ever been was doing 40second anaerobic capacity repeats, based on whatever wko said was my optimized interval at the time. I ended up on the floor barely able to move for several minutes.
It was a very different sensation of fatigue than going to failure on a ramp test, or max vo2 efforts. Each are awful in their own way.
I’m not sure that is how it works. I’ve read somewhere that how one rates the workout difficulty is somewhat adaptive and normalizes you answer. Maybe @eddie can clarify?
Regardless, as long as she’s not failing workouts she will be served more challenging ones, albeit at a slower rate. At least that’s how I understand it.
That’s because simple is extremely effective. The best training plan is one that you can follow consistently and understand the reasoning behind it. You don’t need fancy workouts with a minute at x% and two minutes at y%, you need workouts that target the adaptations you are looking for at that particular moment. Muscular endurance needn’t be complicated, VO2 needn’t be complicated, it’s more about doing the simple things at the correct time and at the correct dosage.
I can’t help but feel, that we’ve allowed apps to overcomplicate training and therefore sell themselves as a necessity, when in fact, for a lot of people they are just an added layer of stress.
Stop obsessing over your exact FTP and just do the work. The system works.
Here is the fact: you are getting stronger, a lot stronger, by simply doing the work.
I see so many people on here overly worried about their exact FTP number. The old system used to make us wait four long weeks of dread just to test and tell us what our FTP was. This new continuous progression system is a massive upgrade and a lot better, but people are still getting caught up in the weeds over single digits.
Look, if you listen to anyone who knows what they’re talking about, once you start getting on the high end of your W/kg, gaining 1 or 2 watts is like absolute gold. Just look at the bike industry right now, people are paying $18,000 for a new frame just to gain 2 watts over the old version. We all love fast gear (I’ll take the free speed from a Zipp 606 combo any day of the week), but those wheels do not turn themselves.
The app makes you stronger and faster. That’s a fact. It rewards the daily grind, the early wake-ups, and the relentless consistency of alternating those hard Saturday group rides with focused interval sessions.
The real proof isn’t a number on a screen. Two years ago, I was showing up to group rides and could barely hang on the back. Now? I’m pushing the pace, closing gaps, and on a good day, I’m the one dropping the group.
Trust the process. The engine is everything.
100%
I love TR. I have been using it for years. I think the general concept is sound. But they have leant far too much into AI predictions that simply don’t help or add up.
I have taken to simply ignoring all the AI guff and stick to doing workouts that I know work for me and are slightly harder than the workouts I did the week before. I don’t trust the AI prediction to know enough about me (I’m old and have a bunch of life stress to deal with) to be accurate.
All I need is a sensible plan that will make some minor corrections along the way if I skip stuff and will progress over time. Stop telling me things that you are just guessing at and which make no difference to me other than being demotivating.
…and it’s still in beta after all. It’s not realistic to expect it to work flawlessly at this point.
What serious company forces ALL customers to use beta software that clearly doesn’t work properly?
Give it to a bunch of people who opt-in to test it and let the rest of us crack on with proven tech.
That’s strange.
The 6-watt drop I’m seeing looks like it came later in the morning after you did your warm-up. 
Just turn it off.