I think it probably will do that for the most part, and maybe it was the point. But the system is user data driven. And so that is gonna involve variability on that - good or bad.
But we are getting on tangents now.
FTPD replaces the purpose of the ramp test. Not necessarily the result. That is an important distinction I would bet TR will start to point out if folks start showing ramp results that deviate from FTPD - which I would also bet would happen regularly if everyone tried to validate with the ramp test.
I don’t have any bone to pick that I know of?
And I don’t have any need/desire for any future prediction from AT in terms of MMP. The biggest things I want are: 1) better classifying of actual work done, and 2) a better incorporation of recovery that incorporates bio markers.
Yes I agree some will still be out of Coggan classic zones. And going back to your original points, I feel like some get caught in the trap of having a false sense of security by seeing precise FTP (estimates) on a regular schedule, and precise difficulty scores on workouts, and precise PLs. Perhaps the biggest takeaway I think many should get from AT and FTP is the opposite, that it’s not about precision it’s about sustainably doing work at achievable and productive levels, with an occasional stretch workout to calibrate your feelings and generate additional stimulus. Just a few random thoughts, along with my earlier “hide the PLs and WLs and only show achievable/productive/etc” comment.
Ah, we are speaking different languages here.
Language 1: FTPD spits out a single number called FTP
Language 2: FTPD spits out multiple numbers which represent total fitness
@jjmc what parameters does FTPD spit out in your vision?
This is spot on in my book.
I guess the first one, but even though it is called FTP it is openly acknowledged by TR as not FTP - at least not how Coggan defines it.
But most importantly, TR intends FTPD as a way to keep users progressing their fitness. In theory, it should be a number close to a ramp test. But in practice it needn’t be and if it doesn’t, that does not mean it is flawed.
Finally had a chance to try this out (missed the launch by 2 days from the last ramp test that was on my calendar) and got a bump of 4%. Clearly was hoping for more but isn’t everyone? Thought I might take a ramp test to compare once I did get chance to try it out but I think it’s good enough. 
New TR user here - quick question on when AI FTP detection will appear in ‘Career’.
Have signed up to TR and synced with Strava (indoor and outdoor workouts with power and HR), but not completed any TR workouts or ramp tests yet.
Have also enabled AI FTP in trainerroad.com > Account > Early Access > AI FTP Detection ‘Enabled’.
When will I see the option in ‘Career’ to use AI FTP detection? i.e. when will my first scheduled ramp test appear in ‘Career’?
You need to build a plan through plan builder etc which will have the first workout as a Ramp.
Or just schedule a Ramp.
Either way it has to be on your calendar before it’ll show up on the Career Page.
I think it appears after at least 12 rides with power.
I did an unscheduled gravel time trial yesterday (Saturday) that was approximately 24 min full gas. Monday, I have a ramp scheduled, which I don’t plan on doing. Will this AI FTP estimate look at an unscheduled outside race/tt to estimate my ftp or will only look at previously scheduled workouts that were completed (inside)?
Was it done with a PM? In any case, no single workout is figured into FTPd. It’s not an algorithm that looks at your best 20+ min recent efforts, there are plenty of programs that do that. FTPd looks at many different factors and estimates what is an appropriate FTP. So it could probably ignore your entire last week of training and FTPd would still be within a few watts if not the same.
Yes. Done with power meter. In the past, when it comes to doing FTP test, I normally sign up for a time trial and use the appropriate multiplier. Something about the competition aspect of the paid event, seems to motivate me to go all out compared to looking at a monitor.
On Monday, I guess I’ll let TR do the estimation. It’ll be interesting to see how that compares to 90-95% of my race/Time Trial power number.
A % of your best 20 min race effort will likely lead to an overestimation. A 20 min FTP test requires some VO2 efforts before, usually at least 5 mins, to deplete anaerobic contributions.
Agree. The actual effort was 24 min. I did some VO2 prior to and the day before I did 2 hours of hard XC riding with 90 min of it in zone 4 heart rate. I definitely wasn’t as fresh as I should have been if I were planning to do my best effort. I was a last minute decision to do the TT.
Just a follow up 1.5 weeks after accepting FTPD that I know is 3-5% high compared to my sustainable power.
Every workout has been right on until today. I had a 2 hour level 8.4 - Garrowby +8. Basically 4x2x10. I popped after the two highest intensity ones, which are 94% of current setting, but certainly higher into my threshold zone.
I dropped to z2 through the last 2.5 10s and simply answered the survey as too intense. Adaptations down in level moving forward, which is exactly what it should do.
So far so good. The system is performing exactly how it should for me, even with the high FTP setting.
Note, I have also skipped workouts and done easier ones. The system handles all my mucking with it perfectly.
Hit the button. Sweet, 2.2% bump. I accepted.
But. Ftp did not change. Do I have to manually change it? Do progression levels update from a manual ftp change?
What is FTPD, is this a TR expression to AI FTP?
It’s a community-defined shorthand for the name of the feature, “FTP Detection” (in the topic for this thread). See also: