Adaptive Training is now in Open Beta!

Forgive me but I disagree.

There is a broad spectrum between Jefferson (4x5@95%, level 1) to Unicorn +1 (60 minutes @95-100%, level 10). The first might feel easy to moderate while the latter probably will feel very hard to all out. Both workouts have their place as not every session is required to be a productive, stretch or even breakthrough.

Also, picking up your example, I think it would make sense to decrease levels after a 5 watt bump from 4 to 2 or even 1. At least as per my experience: Before my last bump I did polar bear +2 which is rated 12.5. After increasing my FTP, my sweetspot level went down to 7.x and this has proven to be just perfect. I certainly couldn’t get away with a 9 or 10. So yea, I guess for some it might work.

However, the absolute decrease shouldn’t be a reason for concern. Your feedback will directly influence how quickly levels increase, how long they stagnate or perhaps eventually decrease. So, again, as per your example, if you rate a prescribed threshold workout as easy, AT will adapt. Maybe not after the first but most certainly after the second “easy” rating.

2 Likes

The issue I have is just from the math perspective and error rate in testing… A 5 watt change on a 250 watt FTP is 2%. It is just within the realm of being the same FTP. Small changes in FTP should have small changes in PLs.

1 Like

I guess that will be the new TR controversy.

Before AT we had folks complain that their ramp test result has put them in a place where they cannot complete workouts. Now people complain because AT decreased levels by “too much” and subsequently fear it will be too easy. Though I guess if people are honest with the feedback, AT will make it hard enough for them in no time.

Personal anecdote, all sweetspot workouts of my low volume base have an IF >0.90. That being said, levels are down by 5 due to last FTP bump.

Every workout is moderate or hard to me. Nothing is easy unless it’s Pettit or less and nothing is more difficult than hard. So I’m always stuck levels 1-4.

I don’t mind lowering progression with ftp increase. But 5W ftp improvement shouldn’t take you from a 3.7 to a 1. Perhaps 3 would be better. If you were doing 15 min sweet spot at 235, it’s not that much harder if at all to do SS at 238

1 Like

You never get progressions high enough. The moment you work your way from 1 to 3.x, the next ramp test comes up and any change, even 2W, lowers back to 1

Do an alternate workout to bump the level. My vo2 has had decay take it from 6 to 1.2. Next vo2 I do will not be starting at a level 1 when I know those in the 4s are generally not difficult for me.

1 Like

Well, now I’m confused.

I understand but The whole point of ML and AI is to adjust to me so I have a steady progression without burning out. That’s what I think the main goal is of AT. If I start gaming it with alternative workouts and manipulating the post workout survey, I feel that it defeats the purpose.

Actually I am happy with TrainNow most. I know how I feel BEFORE the workout and can make it easy or hard based on that. I may turn off AT because it’s giving me progression level anxiety. Lol

1 Like

I suspect at some point it will work better but now it is in beta. It needs a workaround in my view

1 Like

Agree. Needs quite a bit of adjusting. I’m surprised it’s moving along so quickly

That’s not true. I was part of closed beta and have seen a 10 for SS, a 9 for VO2 l, and an 8 for threshold.

Don’t be. :slight_smile:

You are not manipulating anything. You give honest feedback and AT does the rest. If you think it got it wrong, you are able to select an alternate which perhaps AT thinks is a stretch but you know better.

Just because you feel you can do it doesn’t mean you should. Perhaps you can relate to some of my findings:

Before AT I trained 10-15 hours per week. I basically went through an endless cycle of SSBHV1/2 and SuSPB. With AT I am now down to 5-8 hours and I doubt I could get away with much more. Some of the workouts I see in my calendar I have never seen before (eg. Redondo+2). They are a proper challenge. Others like Starlight-2 are not. At least not on paper. Though how it all acts together, it proves to be productive. I am now carrying less fatigue and see myself breaking all my all time records. I am now one step away of also breaking my all time high FTP (4.62 wpkg) which I thought was impossible.

So yea, I would say give it a chance. It might as well surprise you.

10 Likes

That’s not what happened it me, I would also expect that IF your progression levels did drop, you would be offered adaptions from your old levels to new, which you could refuse, an your progression levels would be pulled back up

Its also worth keeping your eye on the game, you are training to improve fitness (and hopefully your FTP increases), if your progression levels are 1.0, who cares as long as you are achieving this

2 Likes

I workout 2-3 hours a day. If I do less biking, I just fill up the rest of the time with other workouts. It never goes down for me. I enjoy working out.

About PL dropping to one. That shouldn’t happen if I’m correct. After my last SSB2 LV block I was around 4,5/5 allround and after I used plan builder for the first time after a new ftp it dropped my levels about 1 to 1,5 points. Not to one.

I see a lot of people thinking that they should get a PL bump after every workout, but imho that’s not how it’s supposed to work. AT is for fine tuning your workload so that no type of workout is too easy, or too hard. The overall progression is already build into the plans.

2 Likes

Just a question on process.
If I have a 90 min sst workout in the plan but I decide to do a 30 minute Zwift race - threshold session - instead, do I match the two workouts together or just delete the sst workout from the plan or just leave it in place but not completed.
Cheers

When I adjusted my FTP (from 272W to 288W) my levels dropped by about 2.0 +/- for SS, Threshold and VO2max. Endurance dropped the same or a little more. All these levels are now around 3.5. I was then offered a whole raft of adaptations, which I accepted. The changes weren’t massive: Pisgah to Saddle Mountain as an example, so a drop in levels of about 1.2 - 1.5. So for me, it seems as if I’m being eased back into things following a moderate FTP increase which is sort of what I’d expect.

Because we are all being offered individualised adaptations it’s hard to generalise/extrapolate from our own experience. Levels dropping from say 7 down to 1 following an FTP increase from 200 to 202W definitely seems wrong and might need a call to support.

3 Likes

Odd thing I noticed with Workout Levels. Any explanation for the problem below?

Had Engeler scheduled, VO2 max 5.0. Six 2-min intervals @117%. With some zone 2 at the end.

Basically, I wanted a shorter version (~1 hour). There weren’t any alternates that were 1 hour in length that matched, and I’ve missed out on progressions in the past by cutting the workout early. So I created a custom version. Literally all I did was shorten the Z2 at the end, that’s it. But somehow, the workout got harder? VO2 max 5.3 (from 5.0). What? I shortened the workout but it got harder?

Then I found this workout. Monadnock -2, VO2 max 5.1. And this has six 2-min intervals @120%. So harder intervals, and even shorter rest, but still rated lower than my “custom” workout. Any ideas?

Seems like something you should contact support about.

1 Like

The workouts in the catalog were graded by excluding the warmup/cooldown sections, and occasional further hand-tweaking where the TR team felt it was necessary. Custom workouts are computed on the entire thing. So in your case, the “shortened” cooldown looks like more work, because the original workout was graded as if it wasn’t there at all. Also, the warmup ramp was likely excluded in Engeler’s computation, but included in yours, so that appears as some extra threshold work.

I haven’t seen anything official about allowing custom workouts to mark the warmup/cooldown so PL calculations are more in line with the catalog (or allow hand-tweaking). I’m guessing that TR is hoping the fix is to throw the “random outdoor rides grading” algorithm at custom workouts, because those will have similar issues around warmups and cooldowns.

5 Likes
  1. Personally, I wouldn’t associate an unstructured ride (such as a Zwift race) with a workout that I happened to have scheduled that day. TR’s workaround they’ve implemented for now means that you’d be given automatic PL credit for the workout (sweetspot, in your case) even though the on-bike work you actually performed (your Zwift race) likely bore no resemblance to that workout.

  2. Up to you whether you delete the uncompleted workout or leave it there, uncompleted - functionally there’ll be no difference in impact on AT. I’d probably leave it there, so that if I looked back to that day in my TR Calendar I’d see both what I’d intended to do (the uncompleted workout) and what I did do (the Zwift race).

1 Like