Thanks for taking a look. I get your point but if my memory serves, I only missed one deload/base Saturday for travel and one workout cut short bc of DOMS from a heavy leg lift day. But this is very rare. Only two examples causing this? And they happened in my Base Phase, so why is my FTP still going way down from the opening prediction in my Build Phase when I haven’t stopped any workouts short in this phase?
I also dont really get how to or why we have to “game the system.” I thought if I ranked my RPE on the slightly easier side it would do it but it doesnt.
Good trip. I’ll do this. future question potentially being: okay I did all the things and RPE match and they still went down on FTP lol. ill report back to you
Week 29 you had only one workout. Week 30 you missed one SS workout and week 31 you missed the endurance ride. And then you cut your threshold workout short the last week and still got a 2% increase.
Week 29 (July 13) was the end of one training plan. started a new one the following week.
New Plan:
Week 30 and 31 - agree. very rarely do I miss workouts but my job got in the way
Threshold - ya I bonked from DOMS from heavy lifting
But irrespective of these three, I can confidently say ive done plenty of training plans on the new AI and didnt miss workout and the same thing still happened. I for sure wouldn’t ask about it if I routinely didnt complete workouts.
This really points to TR needing to create a compliance metric that somehow combines:
Compliance to plan as it was at the start of the 28 day AI FTP Prediction period
Compliance to the AI updated / adjusted plan
Users shouldn’t need to do crazy things to understand why the detected AI FTP differs from the originally Predicted AI FTP. They system should “tell” the user at the end of the prediction period what it “saw” during the previous 28 days that caused the detected AIFTP to differ from the original predicted AIFTP
I should share this alongside that, since this is the TR-approved version..
How to Rate Your Effort
1 - Easy
This ride felt easy and non-taxing, requiring little effort or focus. You could repeat the ride and pass it without issue.
2 - Moderate
This ride was somewhat comfortable but required some focus to complete. You felt a little challenged but had confidence that you could finish. If the ride had an additional set of intervals, you could complete it.
3 - Hard
This ride required effort and focus and was challenging to complete. It felt tough, and you looked forward to the ride ending. If there were an additional set of intervals, you could have done it with significant focus.
4 - Very Hard
This ride was very difficult to complete, and it tested you. If there were one more set of intervals, you wouldn’t have been able to complete it.
5 - Maximum Effort
This ride was extremely difficult. It pushed you well beyond your abilities and took a massive amount of energy and focus to complete. You felt like you barely made it to the end of this ride, and that you had to pull out every mental trick in the book to finish. If you had to pause during a workout to catch your breath or adjust the power demand down, the workout was likely a maximum effort.
If a workout was a maximum effort because of external factors like illness, poor sleep, or fatigue, be sure to add that detail in the follow-up question.
@Incuspy, I just took a look at your career, and I think there are some things that I could clarify and help with. Do you mind if I share them here in this thread?
This is a great idea. A short written review of the 28 day window would clear up a lot of questions users have I would think. “You lowered intensity here and there, missed this workout, rated this workout very hard when hard was the prediction, your ftp was adjusted to xyz for these reasons” etc etc. personally I would find this very insightful
@eddie yes please. This is exactly what I am looking for. I appreciate all others comments but I would love to hear from the TR team. Other than 1-2 missed workouts due to job and one bonk due to DOMS I usually complete 100% of my workouts in training blocks. let me know what you see.
Plus I was thinking deeper analysis: your HR on the work intervals in [insert name of Threshold workout] was 5 beats above your normal, indicating that you either aren’t fully recovering, or this workout was at the top end of your current capabilities - or better analysis
@eddie another example. I just finished my Friday 8/21 threshold ride. did entire thing. rated it 7/10 (Zwift) RPE and it knocked my AI FTP down again. seems like when I err towards higher RPE it lowers AI FTP. but even when I give easy rides a super low RPE it doesnt boost it up. seems so unilateral.
Looking at your recent FTP changes, you had bumps up to 248 in May and 250 in June.
Your July detection stayed at 250, but that’s likely because of the dip in TSS due to travel and the switch to a Rolling Road Race Specialty phase, which contains mostly short, hard efforts. Your fitness was likely still going up, but we didn’t think increasing your FTP would benefit your training.
Since then, you’ve bumped up to 255 on August 9th, and you have a predicted bump to 269 on September 6th.
How are your workouts feeling? From my perspective, it looks like things are going well. I’m not sure where you’ll end up on the 6th, but I imagine you’ll see a bump in FTP.
The symptoms you’re describing sound a lot like the bug we’re working on for Outside Workouts, so I wonder if there is something related that’s causing you to see high predictions that drop as you progress.
As @Pbase stated, I’ve confirmed with engineers that we don’t look at Zwift workouts as outside rides, but it looks like you’re doing all your work exclusively in Zwift. Maybe there’s something there that I/we missed..
The good news is that your workouts appear to be going really well, and your fitness is certainly improving. It’s just the dropping predictions that are a distraction and creating negativity around your progress.
The workouts through a 4 week block can go well, but if they’re easier workouts than what was originally prescribed on Day 1 of the block (because the AI has adapted them based on performance/RPE/fatigue/the moon being in Aquarius) then it’s very likely, if not inevitable, that detection on Day 28 will be lower than what was predicted on Day 1.
Just out of interest, do you guys have a way to virtually roll back an account to a previous date in order to reset everything to ‘as at Day 1’ to look at what workouts were originally prescribed?
Recently had a workout that when rated “very hard”, as opposed to “hard” my FTP fluctuated by 9 points. In my personal opinion, without additional contexts, that’s way too aggressive and gives the impression that the alg is over sensitive. We all know FTP isn’t static and fluctuates quite a bit from day to day, but when we’re locked into a training number for a month at a time, we would hope those fluctuations are being considered. As someone who trains 6 days a week I can attest that there are so many factors that go into how hard a training session is. Weekly volume, time of day, sleep quality, nutrition, temperature, indoor/outdoor, hormones, motivation, etc. Most training sessions, regardless of one’s fitness, will easily go from hard to very hard due to the slight increased heart rate as a result of a 15 degree temperature swing.