TR is not for outside training?

Oh good call. Yea, I’m on a Wahoo and it automatically creates the laps.

Hey guys! Jumping in to offer some clarification on Outside Workouts, AI FTP detection and Workout Levels V2

Completing an Outside Workout

For Outside TrainerRoad Workouts, you have the option to get the Workout prescribed based on power targets or RPE targets.

This is so that athletes who don’t have power meters can still benefit from structured training and Adaptive Training. Every Workout (Inside, Outside, based on RPE or power), and Training Plan is created based on the science of getting faster. If you have a power meter, we recommend following the Workout based on power. This allows you to reflect on the workout objectively and consider whether you followed the Workout as prescribed.

@amitpt, I took a look at your Support ticket and I believe the advice you were given was based on whether you would like the extended warm-up duration to count toward your Endurance Progression Levels. If you consider your extended warm-up as an additional Endurance Workout, then you may wish to separate the ride files and associate the extra Endurance work with the TrainerRoad Endurance Workout that you were following.

Otherwise, if you were simply extending your warm-up, there is no need to create separate ride files.

Reflecting on the Workout

This is solid advice. If you follow this advice, you will reap the benefits of structured training with Outside Workouts. Your Outside Workouts will be considered for Adaptations based on how well you report to have adhered to the Workout prescription. Your Plan will adapt accordingly, and you will receive all of the benefits of Adaptive Training.

Adaptive Training and Outside Workouts

If you indicate that you completed the Workout as prescribed, your Progression Levels will change in accordance with the Workout you completed. For example, generally speaking, if the Workout Level is higher than your current Progression Level in that specific zone, your Progression Level will increase. The subsequent Workouts in your Plan will adapt (if necessary) to ensure that the difficulty of your workouts are progressing at a rate specific to your fitness, as indicated by your Progression Levels.

AI FTP detection

The recent changes made to AI FTP detection mean that you need to complete 10 indoor TrainerRoad Workouts before you can use AI FTP detection. I appreciate how this is an issue for you, @amitpt, and so I have raised this with the product team to see if we can make changes to this.

Future Plans with Workout Levels V2

With all of that said, and as mentioned above, we are currently working on Workout Levels Version 2, which will analyse the power of all Workouts (Inside, Outside, Structured, Unstructured) and adapt your workouts based on this analysis.

I hope this clears things up. Please don’t hesitate to let me know if anyone has any questions about any of this! DM’s are always open!


@SarahLaverty ,
Thank you for addressing the issues !
One request though in regard to your request regarding AI FTP detection:
Please don’t change it back only on my behalf if it is not as correct and precise for outside sessions as it is for inside ones.
Your scientific approach to training with evidence based data is what I’m looking for when I use your training tools.


Thanks Sarah , now i can sleep well :slight_smile: I was a bit concern what i’m gonna do in the spring when outside workouts will come up :slight_smile: THANKS :slight_smile:

So glad that was helpful, guys!

No stress, this will always be our priority. You’re in the right place :slight_smile:

I can add some of my experience from training during this summer season with outdoor workouts only, though I had quite a good spot to train (ride around stadium). Though even then it’s very hard for me to nail the over-unders outside. I never reach the same sensation in legs as indoors. I guess that’s due to my lack of skill to keep steady power compared to ERG trainer and this ruins the goal of the workout.

For other workouts I might also failed to keep them exactly to the book, as my Garmin execution score (almost) never went above 60%. But I take this measure with grain of salt, as it counts both going over as under as non-compliance.

Lastly, I recently came back to indoor training and my FTP took a slight hit, but I can explain that with change of power meter data source (pedals vs trainer), and some of progression levels where set too high.

Anyway - even with that, when I compare my average speed on unstructured rides, I’m just riding faster, so still I managed to get the benefits.

There is no doubt one can get benefits from training with TR outside.
There are, though, two issues:

  1. It is not scientifically accurate as promised and relies heavily upon one’s subjective feelings.
  2. It is not as accurate as indoor training as promised.

It’s manageable.
Maybe the data is not there and it’s OK.
TR is already way above many other online training platforms.
Just don’t claim it is.

I also respect other views and extremely appreciative of TR’s outstanding support.
I only wish I could rely upon TR’s data in order to improve my weaknesses correctly according to hard objective scientific data derived from my own execution compared to planning.

1 Like

I largely agree with the observations. I was completing outdoor sessions all summer, only to find my FTP was decreasing. The AI stated my FTP should of been increasing. I now complete at least 80% indoor workout so TR can work with the data and give an accurate a point of where I am.

How disciplined are you staying in prescribed power range? Too big variability (VI) during intervals or long steady ride will lower actual gains for you. We should not look at whole ride average power or NP, but continuously try to stay in zone. So, if you say to AIFTP that you completed outdoor workout as intended and marked it as Easy or Moderate, it’ll believe you and predict new FTP according to whatever magic. But if you really wondered from prescribed zone to one below or above, your real progression obviously diverges from prediction. Hopefully upcoming Workout Levels v2 will not just take your word for it but actually look into what you did.

For me personally (mostly riding outdoors), AIFTP predictions have been very good. Especially compared to Ramp tests, which tend to over-estimate for me. During Z2 rides VI is within 1.04-1.08; during Z4/Z5 intervals VI=1.

1 Like

Discipline can take only a certain distance when it comes yo validity.
So do average power vs NP.
With outside rides done even at a very proper open public criterium there are small inclines and small declines, changing wind direction, wind gusts, sharp turns that force sharp changes of power, traffic…
and do believe me I am very strict with myself when it comes to doing intervals.
So sometimes my VI is indeed not the best due to outside influence but NP puts me in the right spot, so does RPE.
If TR could analyze such data it would have calculated how much did a certain execution indeed achieve required goals and grade them.
With the coming winter and added outside elements influence the validity of TR’s sessions without a proper scientific analytical tools is going to be even less accurate.
On top of that, with my science degree and knowledge I am by far not a trained coach nor do I specialize in human physiology…eg…my deduction of how good the workout went can never be as accurate as actual analyzed data made by proven equations written by professionals.
On top of that…AI FTP detection is no longer available for outside rides.

Its still available but you have to do 10 indoor TR workouts first. I’m guessing that to put a decent ‘spade in the ground’ to later adjust with indoor or outdoor workout data.

I think you are correct. Staying in the zone was difficult, especially when fatigued. I believe this skewed the result because the AI does not detect nor correlate the zones or the interval.

I think the indoor training sessions are showing my weaknesses and allowing AI to review and adapt the approach. This has been evidenced several times.


If it is scientific purity you are after above all else, then testing in the lab is the way to go. It’s unrealistic to expect crystal clear data analysis or an accurate AI FTP from a small inaccurate dataset of just a few months of outdoor rides under imperfect conditions. AI/machine learning to some extent works better over time with a large and/or “clean” dataset. So I can totally understand why TR engineers are recommending 10 indoor rides. I have the benefit of years of imported data from TrainingPeaks, and I’ve found the AI FTP detection to be exceptionally accurate based on how the subsequent workouts “feel”. But I have enough experience to know when my suggested VO2 max wattage is actually sweet spot due to a poorly executed FTP test. For me, even if I don’t follow any of the plans, the AI FTP detection is worth the subscription price alone!

Like you I’m a mountain biker and train outdoors a lot. I’ve found that more steady-state types of intervals outdoors (long threshold, or 4 or 5 minutes steady at vo2), can be performed relatively accurately outside as hill repeats. The workouts with lots of intervals or precise steps like over-unders, tabatas, ronnestad, or 30/30’s and similar are better quality on the trainer. If you don’t have a trainer or for whatever reason can not or do not want to train indoors, I think that the polarized plans might be a really good option for you, as they seem to have the most steady-state intervals.

Also, it is really difficult to do intervals with friends - at least if your friends are like my friends who either want to race or want to party. Either my intervals suffer or my friends get really annoyed with me. :stuck_out_tongue:


This whole thread is so strange and I think jumping to conclusions based on incomplete data. As a scientist, you might not make so many assumptions and simply ask better questions. And definitely not speak so confidently about a product you only know about from youtube.


I find it difficult to read some of the statements in this thread about ‘science’ as it doesn’t mesh with the science of exercise physiology as I understand it, my coach understands it, and as exercise physiologists discuss it on podcasts.

1 Like

@russell.r.sage ,
I tried to explain myself extensively throughout the thread but I shall give you the facts again:

  1. I have been training XC extensively for a few years now including group and coach including lots of interval work of all sorts (HRT based).
  2. I found out about TR from YT and have been listening to its Thursday podcasts.
  3. In those podcasts the TR team repeatedly stated that TR adaptive training could be done outside just like inside.
  4. I wanted to try TR approach and find out if it benefits me coupled with the proven power meter data.
  5. I have trained with TR for a few months till not long ago (damn Covid).
  6. During that time some numbers, hard data, didn’t make sense or didn’t add up.
    I contacted TR’s support and through their professional and very reliable support learned of TR’s shortcoming when it comes to outside workouts.
  7. This thread came to be because after wanting to go back to training I found out by chance, without any notice, AI FTP detection stopped support for only outside workouts.
    That also raised the concern if the data I got before, after using it in the beginning of each block, was reliable or accurate.

I hope the facts I wrote here are clear enough.

That’s the problem. These aren’t facts. #3 is false. Everything in here seems based on a misunderstanding about #3. And #7 is mostly false and another misunderstanding about #7. So these are hardly facts and your assessments based on your misunderstandings shows that you clearly don’t know what you are talking about.


@russell.r.sage ,
I appreciate your point of view but I shall not negotiate documented facts as #3 is documented in more than one TR YT episode for you to search if you so wish and #7 is documented either through current software demand or through TR’s support emails.
I gave zero assumptions nor opinions other than concerns in my replies to you.
I do hope this discussion shall lead to a better or quicker resolution of the issue at hand for I do have the utmost respect for the TR team and I think they might be the first to crack this issue for our benifit.

I’m not listening to many podcasts these days, and can’t comment on what you heard. However I believe it is easy to get confused about how TR handles inside and outside workouts. I find the podcast has a lot of enthusiastic energy, and since training can be a complex subject, they tend to conclude a discussion by simplifying things. Which can lead to misunderstanding.


There’s no real point continuing this thread. We’ve all tried to help, and we’ve been heard. OP isn’t changing their opinion. The title of the thread states an opinion as if it were fact and leads to emotional responses from other users.