Adaptive Training Closed Beta Update

Follow-up on this (forgive me if it’s been answered).

If I am going for a long outside ride, would creating a “Free Ride” Outside workout be a way to capture that ride for future adaptations? Admittedly, I have not looked to see if “Free Ride” is still an option. I am using the AT Beta I am planning to just add my outside rides on the fly (Garmin Edge 1030 to record) and do not intend the ones I have upcoming to be “structured.” They will be longer road/gravel rides with a friend.

No worries!

There is not yet a way for unstructured rides like that, even if recorded on Garmin, to be accounted for in your Adaptations and Progression Levels. HIGH priority to incorporate. The only Garmin workouts accounted for in AT need to be TR workouts, I wonder if anyone has had success with the ‘Free Ride’ workout though? I think that’s a no.

Going to try to get a better reading from the team about timelines/roadblocks we’re facing/where we’re at for unstructured outside rides.


Thanks Ivy. All good. If the rides are productive it’ll bear out in future planned rides and RPE surveys I am sure. I appreciate the quick reply. So far I’ve had good experience with the AT and TN features. N=1 for what that’s worth.




Kinda given up hope will ever get into the beta now! Still waiting! :joy:


I was recently added to beta and have just put together a little “salvage the summer” plan after a lot of time off the bike this summer. Without even starting the plan (begins 7/13) and just doing a little ad hoc Z2 this week, my progression levels have changed… I mean, that’s cool - BUT…

I would really really like to see my progression level history. Has anybody mentioned that here yet (sorry, I wasn’t in beta for a long time and was not following this thread). Not sure exactly how I’d like to see it, but maybe similar to the PR chart, you could filter on zone and instead of duration on the x-axis, it could be a selected date range

Or on the home page progression screen, just add a date filter and it would change the bar graph to what it was on that date.

I’m not a tech person, so maybe that’s wildly unrealistic to store that data, but I feel like it’s just formulas off stored ride data anyway right? The progression levels are cool, but without being able to go back and see them, I’m not seeing much value in them aside from being a little carrot on a stick.

It’s been discussed, and a couple of users have put together some nice looking graphs of their own PL history. The problem with an official history is that you would essentially need infinite PL history graphs/record lists (or you know from say 150-350W or whatever) because PL’s at different FTP’s are going to be quite different.

1 Like

Dang that’s a good point. I haven’t gotten to see how that works in practice and this didn’t even cross my mind. Welp, carrots on sticks are still something lol. Or using the level to select workouts outside of the plan.

Somebody smarter than me please solve this.

You and me both. Along with plenty of others it seems.

I heard that you need to PM @IvyAudrain multiple times and say “pretty please”. :shushing_face:

(Not to be taken seriously).


I just did a threshold workout ranked at 3.4. I passed it and rated it very hard. The next threshold workout in the same block was going to be 3.7 but AI is suggesting an adaptation to a 2.4 workout. My threshold PL is now 3.4. Why is AI proposing workouts below my PL?

1 Like

Better email support because we are guessing at best here.


We are back to how you rate the workout. It does impact the adaptations.

1 Like

Hypothesis: AT takes “very hard” to mean at the limit of your capabilities. So it is backing off so you don’t crash and burn. Plus a “very hard” workout could easily have been an unsuccessful if you weren’t feeling 100%.

Again, total hypothesis on my part

Response from support:

Thank you for getting in touch with the TrainerRoad Support Team!

This is intended behavior - since you rated the workout as “very hard”, the system then knocks your future workouts down a few notches to make sure that you are not overextending yourself in the future.

This is one of the key features of Adaptive Training whereas in the previous plans if you struggle through a tough workout, the next one only gets even harder. Adaptive Training simply protects you from the possibility of overcooking it.

If “Very Hard” results in next comparable workout being 1.0 below current PL, I need to think carefully about the consequences of rating a workout in addition to how it actually felt. The workout was indeed very hard but doing my next Build phase workout 1.0 lower than my PL, as proposed by AI, feels like easing off too much.

I would appreciate feedback from those who have gone through full Base-Built-Specialty program under AI choosing to trust AI each time. Did it work for you?


Adaptive Training has a sense of humour, of a kind:



I’ve had the similar experience recently, here’s the link ty my post earlier in this thread:

There’s a whole thread about this issue:

I agree with you, as do some others. The survey as it is now, is a problem IMO. It’s not meant to be how it feels (according to TR). It’s meant to be, how was the workout compared to what you or it was “expected” to be.

For example: I have been rating all of my VO2 max workouts as Very Hard. Rarely a Hard, but certainly no Moderate or Easy. Because VO2 max is hard to me. I know it’s going to be a tough workout going in. According to TR, I should be rating these as Moderate. Which sounds crazy. In no way or world would I ever consider a VO2 max workout Moderate. They’re always tough. So my VO2 max adaptations have been lower every time. Similar with Threshold and Over Unders. They are hard, and as such, I’ve never rated any of them as Moderate. So far, the only workouts that get Moderate are Z2 rides like Pettit.

So basically, unknowingly, I have been messing up the algorithm. Because I’ve been rating things based on RPE.


With rating of workouts being not based on RPE but on how difficult I expected them to be I imagine AI adaptations would be more accurate. But then “moderate” should rather be “as expected”, etc. Official guidance on this would probably be helpful as a sticker to the Beta thread…


Yeah, that’s the reason we are pushing so hard for that guidance in the thread linked above. I kicked the spurs as hard as I can (via PM and that thread) and am waiting to see what they can tell us.


Thanks for all that, @mcneese.chad; we’re on the same page here and I’ve posted a comment on the thread you refererence.

1 Like

Still waiting for guidance from the team on that one. Sorry for the delay, we know its a big point of confusion, hold tight!