You responded to me in another thread along these lines, but Nate has said that the responses are not heavily driving the AI - especially if picking between one level up or down, you just need to be consistent with how you rate.
I also tested the rating influence on the last workout I did. The model assumed it would be hard for me, but I chose very hard. It only reduced my future AIFTP by 1 watt and lowered my next hard workout slightly down, but still kept the core interval TIZ workout in tact for that next workout. No other workouts were altered.
I’m on a 40k TT plan, am I correct in assuming that while the ai FTP is anchoring to a PL of around 3 during base and build that it will change when I get to Specialty Phase?
Today TR game me Ericsson, which was nice as it is a fairly entry-level sweetspot workout which I have done half a dozen times before, so a useful comparison.
On the new FTP my HR was 8-10 bpm higher than any time I have done it before (a clear 20bpm higher than one time, although my FTP was definitely too low then), putting it into low threshold (low 170bpm range for me), not sweetspot. TR predicted it was going to be hard and it definitely was that - I still was able to do it because it would make a pretty attainable threshold workout as well.
My theory (I guess not 100% serious!) is as follows:
Since I have not been training for anything bike related for ages, I have only really been doing random endurance and sweetspot workouts which - generally - I am pretty confident I can do/are only moderate PL jumps. I can only assume that the AI has seen my training and thought “this guy needs to do harder work to get faster but isn’t doing any”, and its response is basically to jack up the difficulty of the workouts I have been choosing in order to effectively force me to do the harder work it thinks I need.
So it is picking achievable workouts for me, but deliberately putting them in higher zones.
I might just start choosing tempo workouts instead . I assume this will create a death spiral where the FTP keeps going up… I should try and give one of the unattainable- looking threshold workouts a go next… but they all look pretty intimidating from here.
I’ve seen similar. Regardless of the change you make to FTP, the system has a target wattage in mind. Seems when you manually set an FTP now, you’re basically changing the difficulty, but the target for intervals remains basically the same. I think that’s what your experiment is evidencing here.
This technique works for me - as I manually lower my FTP setting, the predicted fail probability for a pinned w/o (such as your Cahuilla example) also falls: the system “knows” my wattage capabilities, regardless of any manual FTP number manipulation, and hence a w/o that’s defined in “%s of FTP” terms becomes ever easier the lower that FTP number is manipulated vs. my actual wattage capabilities.
Hmm, I just tried again. Strangely enough, I’m getting different numbers than before, but still the same prediction regardless of what I set FTP to. This time, I tried moving it higher and lower, but they still all match in predicted difficulty at dramatically different wattages. @eddie , any thoughts on this one?
I did a little more digging and it looks like it only makes the adjustment for a workout pinned in the calendar, but not if you are looking at an unscheduled workout from the catalog. I’d call that a bug. @eddie , could you please advise (see post above where it doesn’t update with different FTP’s)?
Something odd going on for me. Just put a threshold workout in the calendar and it’s saying I’d find it easy. I beg to differ! 2W below my supposed FTP for 18 minutes
That last workout posted gives probabilities that I would expect for an FTP based on traditional FTP definitions. Is the AI giving you a significantly higher number?
Old AIFTP set me at 295 earlier this month. New AIFTP set me at 318, but I adjusted down to 302. That gives me the type workouts I want (higher workout levels w/ a little more focus on TTE and longer intervals).
That 318 number might meet some people’s definition of FTP, but it doesn’t work for how I use the number and I don’t love the workouts it was proposing at that number (higher intensity with shorter intervals at lower workout levels). The TR estimated difficulty on that 60’ @ 318 threshold workout is very, very optimistic. There is zero chance I’d be able to hold that wattage for 60’ even with a gun to my head, very doubtful I could hold it for 20’. I’m not sure I could hold the 302 for 60’ either, but I could give it a fighting chance. I had over/unders a couple weeks ago with the 1’ overs at 310w and barely completed it (marked “extra hard”). I’m just getting into the season, so could probably knock that out easier now, but pushing 310 is not sustainable for me right now (and certainly not 318).
From what I remember of your posts in the past, I think we’re somewhat similar riders in age and ability.
Like you, I was keeping my FTP lower than TR suggested. But, due to my limited training time, I was packing in more and more time in zone into a limited training window. My numbers were great (for me) but not growing much at all. With this new FTP (bumped about 6%) the intervals are shorter and harder. Threshold may be supra threshold and sweet spot may now be low threshold, but I’m doing the workouts, and the AI (which I very much trust on a relative basis) has me growing 2-3% over the next four weeks. This just was not happening when I was completing all of my densely packed, higher TIZ workouts.
Maybe the projection will ultimately turn out to be wrong (10 days in and still hasn’t lowered the projection), but I certainly don’t feel like I’m getting slower, and at the very least, it’s some novelty after for doing it the other way for years.
Best of luck. I’d love to hear how the training progresses.
Ok that makes sense. So if the AI has me at a mid level PL going into specialty phase, it will leave FTP the same and theoretically I should move higher up in PL and hopefully to holding the FTP longer much like the old system.
Yep, from what I saw being proposed with my higher FTP, the workouts looked doable (and I did the first one). I just have a preference toward longer stuff. I’m lucky enough to be retired these days, so more time on the bike is a positive and I love the long sweet spot stuff. The higher intensity work will come soon enough, I only have the motivation to push a couple of months of that kind of stuff every season. Some of it’s also the racing I focus on, my primary target events are 9-12 hours, so it’s often more about what % of FTP I can hold all day as much as what my FTP is.
Same to you, I’m interested to see how the shift works out for everyone.
I am by no means expert, but if you stick to the sweet spot workouts for a while your body will adapt to the new stress and you will have Lower HR. I have kind of experienced that myself. Also you can choose easier alternative workout. Few month ago I came back from unstructured/low intensity training with lower than usual volume and I felt the same as you. After 6-8 workouts I felt that my fitness is coming back and my HR was lower. Every body is different.
Yes this is something I don’t understand either. The new AI looks only at ‘raw watts’ and not FTP…then how is it assigning me a threshold workout, what is it basing the numbers on if not FTP? Is it that it is coming up with some kind of watt profile for my workout and that is being matched as closely as possible to a workout from the TR library, which happens to be a Threshold workout in this case?
. I assume this will create a death spiral where the FTP keeps going up… I should try and give one of the unattainable- looking threshold workouts a go next… but they all look pretty intimidating from here.
.