Makes some sense if the algorithm thinks doing the Ramp Test would be a sufficient stimulus to potentially move the FTP needle.
Iโd just use the 265.
Makes some sense if the algorithm thinks doing the Ramp Test would be a sufficient stimulus to potentially move the FTP needle.
Iโd just use the 265.
Does it matter when you click the button? Test is up next on my calendar but not until Tuesday.
Will clicking it now vs Tuesday morning give a different number? Not sure if the model factors in rest at all.
My understanding is it is looking at your past workoutsโฆso it will matter to some degree based on what you are doing each day.
@ellotheth Iโm not sure what the original respondent posted as itโs been deleted. Why I chose not to ramp test was multifactorial. I felt the ramp test was not a great representation of my capability since my VO2max is pretty low. VO2 max might be something I could improve slightly with training, but gender and age for me are definite limiters. I did better with the 8 minute test, but that test had its own limitations. Ultimately I kept meaning to do a ramp test when they could come up in my plan, but Iโd always come up with a reason to defer to next time. They are just unpleasant, and for me that unpleasantness didnโt translate to an increase in training benefit. I know Iโm not the target market for TR, but I do enjoy being faster and suffering less than my riding buddies.
not for at least 28 years I want to be called โmasterโ
Iโm not an expert but I think youโre absolutely the target market for TR! Time crunched, want to get faster
![]()
(And I wish I could nail my FTP that well with RPE, thatโs like a superpower)
I am glad you added this new feature.
Yeah Iโd be curious about the AI FTP algorithm as well @ambermalika. When this feature was rolled out it gave me a 1.9% increase, about what I would expect. However, since then Iโve successfully completed workouts resulting in increases in PLs (VO2 +0.2, Threshold +0.7, Tempo +0.7 and Sweetspot +0.9. However when I ran the AI FTP detection today itโs now down to a 1.6% increase. Not sure how doing harder workouts, which I used Alternates and chose above what was scheduled, results in a lesser FTP prediction. The numbers are very close and insignificant overall, but it just seems like odd behavior.
I going into a recovery week and Iโll be curious if it changes at all. Does fatigue/freshness somehow factor into FTP prediction?
So I tried out the new AI FTP Detection feature today and it gave me a 14w increase from 252w to 266w but nothing changed with my PL levels. They are exactly the same as before accepting the new FTP. Wondering if @ambermalika could have a look to see why the levels havenโt changed since the FTP change was significant for me.
Updateโฆlevels have now changed as well. Must have just been slow showing
233 to 245.
A year ago I was at 269 (life time high for this 62 year old) and then tore a calf muscle in CrossFit. I have used TR for three years now.
Took all summer to get to 245, and then late fall I got sick. A week prior to officially discovering I was sick I test at 233. Was out most of Dec. Now six week of base program and consistent workouts where SS felt accomplishable I test at 245. Seems so rational.
Btw, no more CF crap. Tr first and two days a week of basic weights.
@ambermalika or @Nate_Pearson maybe you can address this on the podcast, but really curious how close you are to having unstructured outdoor rides count towards progression levels now.
I have to imagine you have this working in the background, at least partially, for AI FTP to work since its factoring outdoor data.
Obviously calculating how my random coffee ride influences my sweet spot progression is a nontrivial problem to solve, but it seems like youโre close, and will really make the trifecta complete with AI FTP and Adaptive Training.
Youโre good ![]()
We are measuring performance. But instead of one test we look at much much more.
Yes, it will work for you.
You just add a ramp test to your calendar right now and it will adjust.
Thank u. Love TrainerRoad
I tried detection again today, no new rides or workouts since yesterday and today I got 246, 1 Watt less than yesterday
Also waiting on the โprocessing dataโ. When we get that message, what should we do? Hit โgot itโ, close the TR app/not close it? Do anything else (except postin here of course)? After I hity the โgot itโ button, I canโt see anything happening. Might it be an idea to pop up a "processing FTP"message so the user knows something is happening?
I did delete the test and replanned it, but it did not change anything. Been waiting for about half an hour now.
Edit: got a new FTP, just a case of not enough patience. It was somewhat lower (246 => 241) which is actually better than I expected (messing with some medicines, and trying to get the dosing dialed in isnโt good for your training).
Small piece of feedback for the team. I spent a while this morning trying to figure out why I wasnโt getting an option for an estimate when I went to start my ramp test from the calendar like I usually do. I missed the part where it needs to be done from the career page (which in fairness it does say in the opening post of this thread!). I personally always keep the calendar tab open on the iPad version of the TR app and rarely use the other ones, so I would have totally missed how to use the estimation if I didnโt already know about it.
4% FTP boost though which feels about right to me.
It is absolutely wonderful, not having to do an FTP test any more!
How good will AI FTP Detection work for athletes, who include other aerobic activities but cycling in their training calendar? I have a week of intense cross-country skiing at elevation in front of me, which shows as โa week offโ in my trainerroad training calendar. A ramp test is scheduled after my return plus some rest days.
Should I check, what FTP value AI FTP Detection will provide and then actually do the ramp test to compare the results?