It’s not significant. He has a hole in his warmup, not an interval or recovery portion. He did the next 148 minutes perfect. It’s little things like this that is rendering the FTP predictor as a silly gimmick and not anything valuable to me.
5 Likes
On a two and a half hour endurance session, certainly you would not think that a pause within the first minute or two of the warm up for a technical issue would not impact the results the same as a pause 2 hours into the session.
Additionally, and anomaly such as HRM dropping to zero while power and cadence stay the exact same, should be seen as an anomaly and treated differently.
1 Like
I’m not sure if it’s a viable work around but could you raise the intensity of the workout (as opposed to changing the FTP) in order to get it where you think it should be and then rate that effort? If the AI is truly looking at the work, hopefully seeing you do more work and rate it appropriately may steer the AI in the correct direction.
1 Like
Exactly. Thank you. This was as near perfect of a 2.5 hour training session as one could possibly imagine. My hole was within the warm up. I hadn’t even had my first interval yet. I immediately discovered the technical issue and resolved it.
Additionally, the drops of HRM should be seen as an anomaly. My power and cadence remained exactly the same. Heart rate dropping to zero is clearly an indication of a technical issue, and not that my heart stopped beating.
2 Likes
I never said I thought those things have the exact same impact.
Our solution is to do a Ramp Test because all of your recent training data is drastically inaccurate, so it’s as if there is no data at all.
For new athletes with no training data, we recommend ramp tests to get the ball rolling also.
10 Likes
The timing of the drop matters. Pausing the middle of a interval is way different and should be treated as such. I do 4-Hour training sessions on saturdays. Am I not expected to pause and run and go pee? If I do, being told that my session wasn’t completed correctly, and at my FTP is going to drop as a result is just silly.
1 Like
I am 100% with you that the drop in the WL score seems excessive. But another perspective might just be…….don’t worry about it. You and I know you did the work. The benefit is there. The machine is gonna score how the machine is gonna score. But will it have any significant impact on your future training plan??? I don’ t know but would like to give the benefit of the doubt to TR and say probably not. I think Jonathan expressed a sentiment in one of the introductory podcasts that I am following which is to just do the workouts I am given and don’t stress out on WL or any fluctuation in predict FTP as long as I am passing the workouts with expected difficulty and feel like I am making progress.
6 Likes
I do think there should be some way to “reset” the AI’s assessment of fitness. I agree.
But sticking in an FTP you think is right and then finding a level 3 threshold workout isnt difficult.
Nore is doing one or two workouts that are too easy at an FTP that is too low.
They could be sat on their phone complaining on the forum at the same time if they wanted 
Edit: It seems a ramp test is the official corrective action…. sounds fair to me.
2 Likes
I had two toilet breaks during my endurance ride last Sunday. They didn’t affect the level I was assigned for the ride nor my FTP prediction.
I get people not wanting to do a ramp test to re-calibrate their FTP because we have been spoiled with AI FTP for a while now. There used to be a time when we welcomed the ramp test over previous test protocols. The ramp test is likely less than 30 minutes total and some people are saying they don’t want to waste their time yet they spend more time on here complaining about how it should work.
I think if TR support suggests that a ramp test should re-calibrate a users FTP and allow them to move forward with the system then that is not that hard a task to do. Just my opinion obviously
6 Likes
Yeah I agree that if TR says you need to do a test to recalibrate due to bad data, you just need to do a test. It was the “they said they couldn’t help me” that I found surprising and questionable. Turns out there was more to it, as @eddie explained.
7 Likes
Is there a way to move the date for AI FTP prediction? I’d like to have it scheduled for either right before or right after my recovery week, but right now it’s on a Wednesday right in the middle of my build.
What happens if I don’t accept the new FTP on the prediction date? does it just move out by one day at a time? or does it just stop working?
2 Likes
Would it not be equally valid to enter what you ‘think’ your FTP is as a starting point and have the system adapt. I’m sure when doing the rides you could quickly work out if your FTP is set too high/too low.
1 Like
I agree, I would like to do this too. Because of when the new system dropped, I am one workout into a new block before a FTP change. I would like to start the next block on the new FTP rather than doing one workout of the new block on the old FTP and the rest on the new FTP. I realize it doesn’t really matter….just cleaner in my brain.
1 Like
The only issue I see with this is that the system isn’t really looking at your FTP but instead it’s looking at the work. If you have bad data in the system then it has no good data. From what others have said, it almost doesn’t matter what your FTP is set at because it’s prescribing raw watts for X duration based off analyzing the actual work.
2 Likes
I guess its only looking at work you have done, and not looking at a given FTP figure as ‘proxy’.
The old data in the old workouts still permanently affect the AI prescription. I’d probably let delete all my workouts by support. Then try to correct some data with Fitfiletools and reupload. Just an alternative. But of course TR support can’t assume the common user to do such things.
I think the recommendation of the ramp test is that you are showing it what you can do where with a manual input you are just telling it a number that it doesn’t look at anyway.
Yeah, I’m right there with you. I’m not going anywhere, and I’m a huge fan of TR and wish it to be the best it can be. I just find this to be a bug for sure. I train 10 to 15 hours/week…500-700+ TSS/week. I have done this consistently for years. I say this simply to express that my fitness history is long and extensive. Pausing because of a tech issue at the beginning of a ride, or a HRM dropping or taking a quick pee break during a recovery portion of a session certainly isn’t truly impacting my performance in actuality.
1 Like