the issue, esp with TR, is a bad FTP estimate has a cascading effect, and progression levels are no panacea. I’ll contend that ilevels with wko gets athlete differences better, particularly in an area like vo2 work, where someone like me does vo2 work in the 110-115% area for 4min intervals, whereas someone like one of my athletes does 130% for similar intervals. TR, even with progression levels, are so beholden to %FTP that having that messed up just makes things worse
And yet someone like the OP is struggling the training. He’s not on the curve detected by the AI.
I’m glad you are here to be the defender of the AI.
Because, clearly, one can do a better long-form test that doesn’t involve short efforts above threshold.
The fact the TR has to run adaptations immediately after an “ftp” test (they still call the ramp teat an ftp test, it was never designed for that because of the obvious issues with using efforts >FTP to determine FTP) shows that it is not terribly accurate.
You can skip the whole adjust workouts until you are in the right ballpark of your FTP by just doing a proper test in the first place.
This saves you time, stops you from doing workouts that are not optimal and gives you a number you can be confident in from day one.
A great explanation.
I can easily get a ramp test result which I can’t hold for 20min straight later.
Before anything adapts you get demotivated, sick and burned out.
If your test result is on the lower side I can see where what you describe might work.
At least if people don’t give up for other reasons, like being afraid of loosing fitness.
I think every athlete figures this out by paying attention to how they perform relative to “FTP” in workouts of different kinds - eg threshold, VO2max, anaerobic, etc.
Even before I started cycling, I knew I skewed fast twitch / anaerobic. This was validated once I started riding with a power meter. For example, my most recent TR workout was a 9.6 PL anaerobic workout. And I could have done harder. But I can just about hold 90% of my “FTP” (ramp test minus 5% for me) for 60 mins - which according to TR is a sweet spot workout, not even threshold. I don’t think I’d be able to complete 5.0 PL threshold workout.
It really helps me dial in workout intensities knowing how I perform in different kinds of workouts - helping me avoid burnout on sweet spot or threshold workouts, and dialing things up on VO2max or anaerobic.
The ramp test, unless they have changed something, just gives you the FTP value as a set %, it is the same for everyone regardless of their ability to perform above ftp. I don’t believe there is any AI analysis of this going on at all.
Since we know that there is a large variance in people’s ability to perform above ftp, and because this varies throughout the season depending on what type of work the athlete is doing, performing an FTP test that actually tests ftp (not a ramp test) is always going to give someone a better number.
Sure you can wait for PLs to slowly adjust towards an accurate ftp but that is a waste of time when you could be performing appropriate workouts from the start.
The fact that PLs start adjusting things immediately after an ftp test is, to me, an indication that maybe the test is not terrible accurate for everyone.
You can see how large of a variance there is in people’s ability to perform above FTP here:
To paraphrase F. Scott Fitzgerald, actively entertaining two conflicting ideas simultaneously is something we should aspire to.
The ramp-critical guys have a strong argument. Firstly, they find “manual FTP” a very effective way to use TR, empirically showing good results from this method. Also, many of them are experienced cyclists, who can “feel” when a workout intensity and interval duration are mismatched / sub-optimal. They suggest the ramp test serves them poorly, for this reason. This is compelling, coming from experienced people.
While a large majority of TR users are very well served by “ramp + adaptations”, there is a sizeable minority who find it demonstrably sub-optimal. This is a group for whom the ramp test is especially poor at FTP prediction. For majority of TR users, the ramp is adequate and highly convenient / easy, but not for this group. I’d also guess this group skews more experienced, those people being more likely to correctly identify the mis-measurement, and resulting pathologies in the recommended workouts. I think the friendly TR devs have provided the ability to manually set FTP to ensure that this group (or anyone with similar inclination) is accommodated.
There are probably some users who are poorly served by the ramp test, but don’t know it. That’s a pity, and hopefully some advice from the forum or TR can set them on a good path.
It’s not clear to me that this shortcoming justifies discouraging / criticising the ramp test for everyone else though (and I’m not suggesting anyone did). Such a discussion would soon lead us to endless variants of the Trolley problem - Wikipedia. (Which I encourage, for anyone so inclined).
Apologies to the OP @gbonhomme . I’ll terminate my FTP side-track here. Someone mentioned alternate measurement methods, and I wanted to make sure I hadn’t missed a TR feature.
Indeed, it would be ridiculous not to be able to ride at your FTP for at least 20 mins
Respectfully, there are a number of flaws in your thinking here, and I’ve emboldened them; no they don’t, subjective assumptions are effectively random, and there is no assumption based on FTP alone - your PLs Are King, not FTP.
They have defined seperate progression levels for a reason, a training reason. There is no logic in selecting workouts “to see if you can do them” - this is the opposite of effective training, it’s akin to racing.
Your last paragraph is excellent, and I would add that appropriately following planned workouts and ticking the workout survey lets TR “dial in workout intensities” for you, saving everyone a lot of hassle.

Because some of us have been doing this for close to 20 years and KNOW where their theshold is.
I’m sure you do. I’m not questioning that.
But - and I notice we’ve tangled on this for a long time now - FTP just isn’t as important as people think. Not here, not anywhere except TP, and not twenty years ago either I’m afraid.
FTP is, essentially, dead. It exists in our discussions for reasons of popularity and nothing more. See Andy’s opinions for more!

And yet someone like the OP is struggling the training. He’s not on the curve detected by the AI.
This is an illogical leap of faith.
This is now largely irrelevant, the only important thing is whether they get the next appropriate workout, which they will if they do a workout and fill out the survey and are consistently progressing over time.
—————
How many people on this thread are actually using TR and TR plan builder?
There were a few points in my previous comment, which part(s) are you saying are irrelevant?
Hi, the one I contest is one I quoted. There’s a million reasons a newbie struggles with training but most of them are really simple.
The second point is an opinion, you are welcome to although I have no insight into the TR AI so I can’t truly be defending SkyNet
The third, actually I don’t see the relevance so I’m neither contesting nor agreeing, it looks like a red herring as my point is the overstated importance of “true” FTP however you try to find it.

FTP just isn’t as important as people think.
It not important to me as I dont use it to decide the power targets during training.
It is of some importance to platforms that base their sessions off a precentage of it. Mainly Threshold and over-unders and no PL levels dont address this as I have pointed out in detail previously.
Sorry I am a bit confused.
The quoted portion isn’t mine, I was just wondering what you were saying is now largely irrelevant as my post touched on a few things.

It is of some importance to platforms that base their sessions off a precentage of it.
It is a factor, but not the only factor, for TR. The most important thing is the PL of the last workout you completed or attempted, plus how you performed and how you responded to the survey.
In time I see no reason we will even see FTP in the app.

Respectfully, there are a number of flaws in your thinking here, and I’ve emboldened them; no they don’t, subjective assumptions are effectively random, and there is no assumption based on FTP alone - your PLs Are King, not FTP.
If I’m following correctly, a few responses as follows.
-
OK. Maybe not every athlete. But many who pay attention to how they respond to training of different types get a pretty good understanding of where they are differentially stronger vs not.
-
Not random. Do you think Usain Bolt skews fast or slow twitch? It’s possible to make inferences based on observed performance. If you saw me play rugby back in my younger days, it would be pretty straightforward to peg me as skewing fast twitch.
-
The OP’s point at the outset was about challenges keeping up with training. I was in this exact same boat with TR a few years back because my FTP was set too too high based on what I did in the ramp test. If the OP is in a similar situation to me, it could be that the root cause of workouts being too difficult is that his FTP is set too high. The TR adaptive training is clearly not working “fast enough” for the OP - if it was, he wouldn’t have posted.

In time I see no reason we will even see FTP in the app.
Amen to that. Agree, hopefully it will disappear.
Id like to see individualisation for ones power curve, and dynamic intervals (power/duration, which is how I mainly train anyway)
Lets leave it there, or move it to the what Id like to see thread as its off topic here really.
To update everyone on this thread, I did the 20 min FTP test on TR this morning. I ate a bowl of muesli and a croissant and prepared 750ml of High Carb drink from Skratch.
Do my dismay, result was 135W (previous FTP was 213W). I felt exhausted throughout the test, with heavy legs. HR was around 155bpm (my max is around 170 bpm and RHR is around 48 bpm).
Was my training too hard for me and I am now exhausted?