Polarized Adaptations from Base to Build

I’ve recently completed a Polarized base block in which I gained 4% on my 20 minute FTP test. Now I’ve started a Polarized build and my progressions seem to have reset at a lower level. For example, the last Vo2 workout I did in the base block was at 6.8 (Barranca) a 4x4:00 workout at 114%. The build program put Saddle Mountain on my calendar for my first Vo2 workout of the block and its a 4.4 - 3x4:00 at 106%. I completed it in standard mode and exceeded the prescribed level by 30 watts or so and my plan did not adapt any of my future Vo2 sessions. Anyone seen similar? Any explanations?

Did you put them on the calendar through plan builder and what did you rate the workout in your post workout survey?

I have the plan set up properly for adaptations as I’ve been getting them throughout the base period and have had an endurance adaptation in build. I rated the workout as “hard”.

You probably didn’t get adaptations because you rated it hard. You have to keep in mind that you are rating the workout as prescribed and not the workout that you did. So because you bumped up the power target you did a completely different workout. I would maybe go back and rate it as ‘Moderate’ or something and see if that changes things.


You are sort of side stepping the system here then wondering why it doesn’t react the way that you think it should. Imagine if you’d done this with a real life coach:

“I’m struggling”
“Well the last three workouts I’ve bumped up the intensity by 10%”
“Why didn’t you tell me? I could have adjusted things.”

Progressions will reset after an FTP test, don’t worry, that’s what’s meant to happen. If your FTP went up then the system needs to adjust things so you don’t burn out.

Either do the workouts as prescribed or retest.

1 Like

You’re missing the point. The workout was too easy. I was doing workouts at 6.2 before my last test and my FTP has increased. Then I exceeded the prescribed workout.

  • And that is not a great thing to do at present. Currently, TR AT does not do a great job of “rewarding” riders for exceeding power targets. I tested this way back in summer with poor results.

  • As of now, if you know that you want a more challenging workout, the best solution is to use the “Alternates” option to find one based upon the scheduled workout, then use the “Swap” function to select that harder workout and perform it at the prescribed power targets.

  • It would be great if TR AT would recognize these manually altered workouts, and maybe it will some day. But we have to handle these differently in order for AT to make the most from our inputs.

Interesting. After being a loyal user for a long time, I left about a year ago and came back to give the AT a try. I love the guys, but it really seems like they have over promised a bit. If it doesn’t adapt to exceeding targets, what DOES it adapt to? Seems more like a programmed progression than true machine learning.

  • As I said, a workout performed around the prescribed power targets.
  • I think there is confusion about where the AI/ML is directly vs indirectly used within TR and for each user.

  • I have no intimate knowledge, but most of the actual AI/ML is well behind the scenes and more about the basic plan progression and setting adaptation guidelines.

  • I don’t think that processing is applied at the finite level of an individual workout. As such, there are limits to what we can expect from those workouts, and how they will apply. We can’t double the power and duration and expect it to adjust to that.

  • Much like a regular coach/rider relationship, I think we are expected to “follow the plan” within reason and the system will update appropriately. If we throw curve balls at it, the system may not handle them as we would expect or like.

  • I agree, communication around this has been fuzzy, with many people (TR and users alike) taking the best possible picture from the info, when there is more nuance and less magic than we might like to imagine.

In one of the early AT related podcasts, possibly the one introducing it, Nate mentioned “super passes”, i.e. you over achieved in relation to the target power. Sometime later, either in a podcast or in response to questioning on here, it was mentioned that “super pass” wasn’t working yet.

I suppose if you are using your trainer in ERG mode then it’s quite hard to get a super pass since the control system will keep trying to bring your power back down to the target.

A human coach would be able to say to you: “do the prescribed workout but knock the intensity back 5%” or “do an extra interval”, etc. AT can’t do that, the only adaptation it can (currently) make is to swap out workouts for something similar where similar means targeting the same progression level.

Oh, and I didn’t miss the point, I responded to what you wrote and explained why AT didn’t do what you thought it would do. (Edit: I only just remembered the superpass bit after Chad’s post)

On podcast 341 I heard Jonathan use the word magic to describe AT.

Right now AT appears to be adapting workouts based on:

  • simple pass/fail
  • post-workout feedback
  • plan goals
  • some TR coaching logic

In the future it may do something else.

Confusing because the original poster wasn’t struggling. The TR POL plan took classic VO2max workout from

  • 4x4-min at 114%

down to

  • 3x4-min at 106%

The TR ‘computer coach’ doesn’t explain its reasoning. To my eyes that drop makes no sense. However the POL plans in general look uninspired to my eyes. :man_shrugging:

1 Like

@Chaser In a quick read through of this thread I didn’t catch any earlier comments, about two things:

First, (and I’m open to correction) my understanding is that since POL is experimental, it gives adaptations to work all energy systems, not keeping the focus on easy/endurance and hard/VO2max and longer SS.

Second, PLs when you update your FTP. Increase in FTP leads to a lower PL; if you think about it, keeping PL at a higher FTP “may” result in workouts that are too hard (although without knowing how AT works it’s hard to know how incrementally they change PL). I don’t have any personal experience with lowering FTP and seeing PLs increase – although my gut (literally) tells me that my next test will see my FTP drop a bit so I’ll likely have an n=1 to report.

  • I’d have to listen back to that cast and context to comment specifically on it.

  • As I alluded to and will state more directly, I think that there is a fair bit of “marketing talk” in play from TR, along with our “greater expectations” as users, that is leading to some of these unmet goals, hopes and such.

1 Like

Don’t bother trying to commenting in context. A few days after listening to just one question on that podcast, all I remember is him saying is magic. Not “its magic under these specific circumstances” but rather “its magic.”

So, simple connection between two words used by two different people, and ignore the context of either. Got it.

1 Like

And so is mischaracterization without proper context.

It was an analogy :man_shrugging: :wink:

From previous posts I believe you have a coach, if you’d had that conversation with him/her how do you think they’d view you…

I agree there’s a lot a hype over AT, in my view it’s a long way from the finished article, but people are choosing to believe that hype. They might wish to consider this flying pig I have for sale a steal at $1m :rofl:

As mentioned above, in this specific range of questions, it’s important to remember 2 key things related to the POL side of this post:

  1. TR POL plans are clearly listed in the “experimental” status. Meaning they are not at the same level of review as their other plans.

  2. TR POL plans are NOT guaranteed to maintain POL attributes with Adaptations, even though we can apply POL to the AT experience. There is not a locked connection to maintain the POL objectives at this time.

And as covered as a separate issue, over-performing workouts to a large degree is not well recognized or handled by AT at this time. The ‘Super Pass’ is something that is not concrete as this time.

So, take those into consideration and apply the recommendations above for a better experience within AT.

1 Like

Yes, but you gave AT mixed signals. Like others have said, it doesn’t take overachieving into account very well. Then you rated the workout as Hard.

I would propose the following: estimate how much harder of a workout you want. Then pick an alternate workout with the desired difficulty. You might need to select the workout from the Stretch category. Finish that workout as prescribed. If that feels easy, rinse and repeat.


Think of it this way… TR wanted to create a plan that aligned with Seiler stated polarized training principles. One of those basic principles is to make hard days really hard, and easy days really easy. Which is why I find going from 4x4-min at 114% down to 3x4-min at 106% to be odd. I don’t know if that is an AT adjustment, or the plan template. If its in the plan template, is there an unstated assumption that your FTP increased from base to build? Regardless it doesn’t appear to be aligned with polarized training principles.

Maybe what you suggest will work. It seems somewhat fiddly when you can just do max repeatable efforts, the way many research studies have participants do them. I heard Jonathan recently say on a podcast that TR wants to collect data on the polarized plans, so perhaps its best to ask support@trainerroad.com for guidance.