As Amber said on the podcast, pros go their entire career trying to nail the perfectly paced hour. Check out all the stuff that goes into the hour record.
It’s super hard to do and takes maximum mental effort and probably a taper to get it right.
Imagine if we told everyone they had to do it once every 4 weeks!
Plus, not many pros that I’m aware of use this method to get their FTP for the reasons I said above. I’m sure some do but it’s not usually a “once every 4 week” sort of measurement.
I would not say “plenty”.
And if we’re going for “close”, why not use an easier or more repeatable method?
Our ML approach is basically doing the #6 on the 7 deadly sins; looking at what you can repeatedly do in training. Then we get your FTP, and AT takes you from there.
But of course, you can still do the hour test if you want! 
We’re not taking away anything, just making it easier for people who HATE to test. And I know there are plenty of people who hate to test.
Some more data for ya’ll, we have about 30-40% of subscribers who won’t test in any form (or are testing outside of our visibility).
They have a worse experience and worse gains with TrainerRoad.
If we could get them to the appropriate levels without them having to do anything…imagine how many more people we could make even faster! 
agree, and LOL not sure I would say “pros go their entire career trying to nail the perfectly paced hour” as it implies plenty. Just injecting a little humor, and I literally laughed at both of those statements. Not at you guys, at how much fun it is to discuss things on forums 
Good stuff 
should be welcome by a lot of users! And will make for some ‘fun’ forum discussions, no doubt!
To be honest, a couple times a week doing two ~1-min Strava segments and going for the KOM (on my commute home). And it worked.
Yeah, should have said “some pros”. But just saying, it’s not easy to nail an hour all out.
speaking from experience I totally agree. And I don’t try them often.
I always look at things thru the lens of my sister. I’d NEVER ask her to do a long test. Ramp and TR’s take on #6 of the 7 deadly sins sounds like a perfect solution for her (she is restarting a TR plan soon…)
I’m one of the folks who loves the idea of NO ramp test. I dread it and would much rather focus on the training. I’m just a Dad who’s trying to have enough fitness to “hang with the kids” while I’m coaching HS MTB-ing. And I do think with machine learning or AI TrainerRoad should be able to estimate my FTP well enough for me to make sure the work I’m doing is improving my fitness. Keep it up TR!
Hey, So my FTP went to 291 to exact and my SS PL dropped to 1.6. Does that still match what you’d expect?
Yeah I see what you were getting at now. This does seem to be how TR currently does it with regards to PL’s/Adaptive. I think the new feature they are talking about (FTP Estimation) will be based on more personal data, as Amber referenced in the podcast from Jan 13th. Not sure if you have listened to it yet, but she mentions that it’s more based on the rider’s performance. When I commented ‘The slopes vary for every rider’ I was referring to how it actually plays out in real life, not how the current PL system works. I think after reading through your reply though, I see what you were getting at and that currenlty TR is using a more ‘one size fits all’ approach with PL’s.
I appreciated your response!
Hopefully it gets released in beta before Tuesday! 
I think this will be a great feature and probably not super difficult for TrainerRoad to estimate with all their data. After so many years of training with power, i can usually guesstimate pretty close where my FTP is before completing a test.
The levels drop so that you should be close to handling it. It is ATs best guess at what your current fitness can handle.
I’ve done this type of thing and the drop by AT is pretty darn close. It has nothing to do with FTP increasing or decreasing, but rather AT accommodating uncertainty in an FTP estimate.
That doesn’t mean there is no increase or decrease in FTP, but that certainly isn’t inherent to the PL changes. It is hoped.
Interestingly, I’ve gotten decreases in PL with manual raises, but not increases with manual decreases. Even when I’ve reset back to an FTP very recently to a point where I was doing 9+ SS at that FTP input. The system is way more conservative with increases than decreases.
At least for me, 9-10 SS PL means my FTP input is close. But still too high unless I do long threshold efforts. I think SS base has worked best for me with a slightly inflated FTP input.
Define worse gains and worse experience. By gains you mean increase in maximal ramp power or repeatability/TTE at different power numbers? The latter is what really matters, and the ramp can indicate that, but doesn’t always.
At some point, if a user is consistent, changes in fitness are gonna likely be less than the ramp test will pick up on. At that point it loses utility and can become frustrating versus just doing something else. I’d guess a decent portion of that 30-40% falls into that category. Not all, but a good amount.
Stated another way, the reason you are interpreting people who do the ramp as being more successful could very well be they are in a rapid rate of growth that positively reinforces the ramp test. So they keep doing them until they stop getting that positive reinforcement from it and start doing other things in its place. The lack of “gains” needs to be carefully assessed here. And, unfortunately, they are also much harder to come by.
This is where PL shines. It gives a more granular measure of fitness increases to track and chase, and should be the focal point of fitness assessment rather than the ramp.
The ramp is good as an initial guess at where you should set training levels - be it new to structured training or coming off a break/inconsistent period, and a good exercise in just going all out. But it is not good as a fitness assessment for consistent users over a long duration of training.
I thought the exact same thing and it immediately brought to mind a Seiler podcast (Fast Talk perhaps) where he was talking about 6 min test being a proxy for MAP and the 60 min test for FTP. His comment was that essentially we need to be able to do those tests and go al-out if we’re serious about understanding and improving our training, or words to that effect. Similar in a way to what I remember KM saying about his testing and TTE protocol, especially that his experiences were that athletes found that easier and more enjoyable as a session,
I can’t quite put it in words but theres thoughts flying around my head about these estimates from sub-maximal efforts & peope not testing etc, all being part of people wanting maximal gains without putting the work in or expecting the big gains without doing the work. There’s also something about the fact that if you cant put the work in during training then how are you going to be able to do it on race/event day? I dont get why the idea of doing a 60 min threshold effort has become seen as some ‘impossible’ or ‘avoidable/unnecessary’ part of training?
I get the point about pacing etc and pros dont do hour tests etc etc but they sure as hell do go to the limit and find those limits in training, so they know what they are capable of when it comes to throwing down.
I don’t know - all seems an odd way to go about training and preparing for races and events to me.
If someone is new to cycling or structured training then you definitely need a way to get them to understand what all our efforts really are. That could/should come from racing or group rides. If not, then you can definitely get that from the trainer. Although maybe if that person doesn’t go all out on races or group rides they don’t really care that much and it doesn’t matter?
Also, a person should be going through the full cycle that includes specialty once or twice a year. Those include race simulation type efforts that should be very hard to all out.
If you know what it takes to go deep, then you know and doing that every 4-6 weeks isn’t going to make you any better at it. You certainly can if it floats your boat. But it isn’t needed. You can save those all out efforts for when they matter, do a handful of very hard efforts regularly to keep an eye on where that point is, and spend the rest of the time mixing hard, moderate and easy during normal training week cycles.
I’ve done plenty of starting to black out, puking after an all out effort stuff. Going to that place too much starts to make me want to do it less, not more. I can do 97-99% of that and have a pretty good idea of where that truly awful point is. And I throw in stretch/breakthrough TR workouts to check on things towards the middle/end of a cycle if my confidence is building through a block. That instantly resets the PLs and keeps things on track.
I guess we should also keep in mind why folks train. At least in my bubble it’s not to excel in racing. Most, don’t race. Neither do I. Group rides aren’t really a thing either. Only exception being bike holidays.
Most folks I know like to ride their bike and enjoy staying fit. Following a predefined structure helps them to stay engaged and motivated. It’s basically a lifestyle.
Doing hourly max efforts or what not certainly wouldn’t do much for that. It just wouldn’t fit with life. Sleep, stress, time etc are all factors. Hence testing should be as easy and straightforward as possible. A ramp test or even better a guesstimate is exactly that. I doubt it makes us any slower.
For sure on the last statement. There is nothing inherently bad about doing ramp tests. There is nothing inherently bad about not doing them.
I have done ramp tests, 8 minute tests and even 20 minute tests. They all give me anxiety and the longer the test the more rest I need and also takes way more motivation, similar to a race. i have also skipped testing and will raise my FTP when I see significant changes in handling long 100% threshold efforts, looking at heart rate and RPE. I end up self adjusting around 2% which isn’t much but it takes away the anxiety of testing.
My garmin eftp is always within 5 watts of what I have my real ftp set too. Anyways, Overall I’m looking forward to the TR version of a FTP predictor which may or may not change how I adjust my ftp. It all comes down to how my body changes in response to those long 10+ minute repeat threshold efforts. Also appreciate the conversation about FTP tests on this forum, great info!
Your point is what many miss including Nate. We are not all the same with the same intentions. Many of us dont race. Some want to push their numbers. Some just want to keep what they have.
Yes I want to be faster but in reality I am happy at this stage just keeping the same numbers. My FTP isnt changing much but my ability to ride outside longer at a steady pace keeps improving.
If I do plans on Trainer road I would only have a max effort on a ramp test. If I dont do that test I dont have max efforts. I noticed this greatly when I was skipping FTP tests and adjusting my FTP using intervals.icu.
Personally I find I improve more when I ride outside and do max efforts…even if it is just to see how much faster I can do a hill or a segment.
I do think a max effort though will get us a better estimate of FTP. Mathematically there will be a model that can use TR data to adjust the FTP. It will be accurate within a range and of course it will not work for everyone.
Yep, your increase was 11.1% instead of 10%, dividing by the gradient gives a new PL of 1.7. The PL only has one decimal place, so consider that 0.1 delta rounding. Did not hear the discussion but not surprised about the ‘FTP estimate’ feature as the ramp-test result really ads little value in the AT framework. It is more beneficial to have users iterate through a wide range of PLs to maximize training variation.
Would not call it a guess, it is a simple linear system to scale PL such that total kJs per workout (of the same duration) roughly remain constant before and after an FTP change.
Yes, described that in the post below and seems to makes sense. Suppose you are struggling to complete workouts at a certain FTP and PL and want to reduce intensity. If you reduce FTP manually but AT would simultaneously increase your PLs and effectively keep you at the same total kJ per workout that would be pretty frustrating and lead to many complaints.
well, we’re arguing semantics. I would say a linear model is indeed the best guess AT has (ok, only guess, but TR model for it). And it has nothing to explicitly do with an FTP change, though an FTP change would correlates with it.
In other words, increasing PL does not mean an increase in FTP. I don’t doubt you can get the correlations you are describing by changing FTP manually. But my sense in using the system and experimenting with it is that you are oversimplifying how it works (and I’m not saying it’s super complicated AI magic either).
No. If there is a linear correlation implemented between FTP and PL, then lowering FTP should raise PL, just like raising FTP should lower it.