Yes it predicted an ftp of 263 (currently at 280). 288 seems very close to where I should be based on my current training
Actually @Nate_Pearson I retract that statement of it being wildly inaccurate. I was feeling spritely today and figured I’d try the workout I always do to start a new ftp progression.
I did Donner and succeeded. Quite impressed actually that I finished it. My brain was telling me it was wrong but sure enough my legs finished it using the air ftp.
You’re confidence in the podcast swayed me to go ahead and give it a shot
I’m having the problem as well.
While I don’t really dread taking a ramp test, using AI FTP Detection makes it possible to use that energy to perform a more productive workout. Besides that, you guys didn’t work on it so hard for so long for me not to use it 
I just checked it, and it’s spot on! Exactly what I was expecting. Thanks for the awesome new feature!
Apologies if this has been asked before but I don’t have the time to go through the whole thread.
Can you just add a ramp test to your calendar in order to use the FTP detection or must it be a scheduled ramp test only?
I don’t think Id change it unless it was way off but I am curious all the same.
Thanks to any and all with the answer
This looks like a great feature, looking forward to testing it.
I think, however, I found a small bug – where the FTP estimated is not properly recalculated when I change the calendar.
- I add Ramp Tests for today and tomorrow. The estimated FTP for today shows 265 W.
- I remove the Ramp Test for today. The estimated FTP for tomorrow then shows 270 W.
- I delete the Ramp Test for tomorrow and add it again. The estimated FTP for tomorrow shows 265 W.
I would expect that 2 and 3 should be the same because both calendars are the same. Also, I am a bit surprised that 1 and 2/3 are that different.
I guess the reason is that deleting the ramp test (or any workout?) does not trigger a reestimation of the FTP for some reason. Also, look like you estimate the FTP in the background for all upcoming FTP Tests, and not just today’s test?
I’m going to leave aside the reason you’re asking this of @StephBee, specifically, out of the ~200 people in this thread excited to not have to do ramps anymore, and instead ask why on earth any recreational cyclist would do a ramp, ever, if they could get to within 3% just by RPE?
Yes. If you have another workout scheduled for the day you’ll have to temporarily move it to another day so the proxy ramp is 1st up in your career.
It was pretty close for me - I had a Ramp test planned for today but gave the AI FTP detection a go mid week and it estimated 232W (up from 225W). Did the ramp test today and came out with 234W. So within 1%. and to be fair I kind of wanted to beat the AI estimate…
Will consider using this in lieu of the ramp test going forward!
Great work!
I use adaptive training but no plan. Will ai ftp be an option for me?I do whatever workout is recommended
If you schedule a Ramp Test on the calendar AI FTP detection will work, you don’t need a plan just 12+ previous workouts/rides with a power meter.
The only thing that will tell you what you are capable of is your own body.
Don’t let artificial intelligence stop you from achieving your potential or be a limiting factor.
As cyclists we are so much more than an FTP number.
Yes you can. If you put in a ramp test on today and wait a few minutes and restart the app you can check it out.
Weird I tried FTP detection yesterday and the day it came out (last Wednesday ?) both times it said 239W (down from my ramp test January 30th which gave me 242) but today it suddenly gave me 247 in spite of not having done any workouts since I tried FTP detection yesterday
Maybe if you hadn’t seen the estimate you may have got higher?
Not a criticism…but as soon as people have s number in mind it can become a limiting factor. The whole ramp test is flawed because of this (it’s a flawed measure of testing anyway) because everyone knows the time of the test that you get an FTP increase, so when people get to this point of the ramp test, RPE naturally increases or people back off as they know they have increased. I’d love to see the data but looking at my own ramp tests vs the old 20 min tests my progress was lower. So I scrapped the ramp tests. Sure FTP estimate will be pretty close, but I reckon we could all get close anyway. The only way to measure performance is through performance.
I marked my session as Very Hard. My Sweet Spot PL actually increased afterwards.
Anyway, it makes sense that the Goddard workout might not be the best case for judging the new FTP. Looking ahead, my upcoming scheduled “productive” workouts look doable. Not sure about a number of “stretch” workouts I have coming up, but hopefully if I start crashing and burning on those, AT will dial things back.
(Why did Plan Builder give me so many stretch workouts, though? I just heard Jonathan say on the podcast that it’s rare / unexpected for that to happen, but it’s common in my calendar. Maybe it’s because I’m on a low volume plan?)
Do you rate your workouts easy or moderate afterwards?
I could hazard guesses, but I don’t actually know. At this point you might want to reach out to support to see if either a) they have an explanation, or b) it turns out to be a bug.
No, if I struggle and take extra breaks but complete the workout, I rate them Very Hard. They show as “too intense” in my history afterwards.
As a long term ramp test struggler I really love this feature, thank you very much TR!
It’s given me a 5W increase which I’m pleased with and think is about right, but I have a slight worry that it might be taking into account erg mode spikes that occur semi-regularly on my Wahoo Kickr. It will sometimes register 100s of watts above anything I could do in real life for a handful of seconds. I hope this kind of erroneous data is filtered out somehow and the increase isn’t being artificially inflated 
(I’ve had a look on the FAQ but can’t see anything mentioned)