Brutal. Thanks for the response. Any idea which podcast?
1 Like
This may not be the latest, but is connected:
But is a repeat of what is already linked above:
1 Like
Did you read what Nate said as meaning that they’d apply WLV2 to TR workouts, done outside or inside? Is he talking about indoor TR workouts where ERG is turned off?
If you are asking me, I don’t think I understand or have confidence to answer that.
- The one exception is that I don’t think trainer mode (ERG or otherwise) is relevant or considered, WRT to inside workouts at least.
2 Likes
I will put my pessimistic hat on: I don’t think TR will ever be able to release this feature as the underlying Workout Levels / Progression Levels are divorced from physiological markers.
Here’s my thinking using a simple example:
- Ride to rate = 4x10 at 80% of FTP (Tempo intervals), but the intervals are done after doing 2 hours at 60% of FTP
- From a workout level perspective, how are the Tempo intervals rated? The same as if the ride was 4x10 @ 80% of FTP, and then 2 hours at 60% of FTP (endurance after the tempo intervals)?
Without a physiological model for WL, how does TR compare the two impact of the tempo sets for the two hypothetical workouts (i.e., long endurance before or after the tempo intervals)? Is 4x10 Tempo after a long endurance equivalent to 4x15 tempo done fresh? Equivalent to 4x20 fresh? How do you figure this out?
3 Likes
This is the latest! Thanks for sharing, @mcneese.chad 
.
1 Like
This may change as we continue to make progress but I believe Nate was simply saying that we would apply WLV2 to all TrainerRoad Workouts before applying it to unstructured outside Workouts.
Correct 
.
2 Likes
For everyone’s sake, I hope that pieces like this can get figured out and it doesn’t just die as a potential feature.
In response to your hypothetical, for me at least, I wouldn’t need it weighted. Treat that like 2 workouts and show me the new levels based the same as if I just hammered out 2 workouts in a day one after another. Let RLGL show you after if it was too much or you shouldn’t have done that second workout. Again though my preference, and the question is valid from data perspective.
Right now, I am pretty content with Garmin’s version of things so I use that for what kind of intensity to do which doesn’t line up pretty closely with RLGL and then come over here to TR to choose a workout. Garmin will also give me an execution score out of 100 which is nice and I wish that TR would add, but would be better to break it out per interval so I can see clearly where and by how much I was off. I’m usually good on the intensity and then off on the recovery is my guess.
My guess, part of the issue is intent vs execution when it comes to outdoor work. I ride MTB, so I don’t have much intent on fun days but it’s still a workout. Garmin just says to take a day off or do an endurance ride most of the time.
While I cannot tell you how this works (because that would be giving away secret sauce 
), I can tell you that Workouts will be considered in their entirety. So the two Workouts you described above would not be “rated” similarly.
As I mentioned, unfortunately I can’t explain how we will measure the difficulty of Workotus with WLv2 but you’ll see the differences relfected in Workout Levels and your Progression Levels.
So no need for your pessimistc hat . We’re making sure to ask all of the necessary questions to deliver a feature that will improve your training.
7 Likes
Agree that PL as designed for indoor rides just don’t seem compatible with unstructured rides.
Most of my rides are outdoors, the problem/challenge here is quantifying the effort due to unpredictable/uncontrollable variables.
I try and match up post ride, but I feel this is somewhat futile.
In my situation I “simply” want TrainerRoad to take into consideration my outdoor effort as part of my adaptive training plan.
I’m not concerned if my Sweet Spot PL is a 4.2 or 7.6
3 Likes
@AlphaDogCycling certainly has a point. Lets forget for a moment WLv2 and see how current workout estimator works: created 2 Sweetspot workouts in TrainingPeaks and synced to TR calendar (SS - Sweetspot at 90%, AE - Z2 at 60%):
Note the Sweetspot progression level differences, even though SS part is same. Thing is, I do first kind of workouts regularly, but not sure I am ready to tackle single-focused Sweetspot workouts at PL 11+ just yet. It would be ~ SS/6x20 a la Marcus Baker from official workout library.
Or in other words, I am actually glad not getting PL boost 
3 Likes
In this example I believe the intended design would be to schedule a 2 hour SS Workout within your 5 hour ride.
Not include the entire 5 hours as a workout.
I used to try and match up my 3+ hour outside rides with a workout post ride. But it is not compatible.
Now I usually schedule a 1.5 hour workout on weekends and usually complete at the beginning of a 2.5-3.5 hour ride
I don’t think it is so simple. There are workouts where you are asked to e. g. first ride x hours in Z2 and then do your intervals. The endurance work induces fatigue and you train to do, say, sweet spot efforts in a fatigued state. So I think it is important to count the entire 5-hour ride as one workout.
2 Likes
Point was slightly different: if possible WLv2 estimates external outdoor workouts in same way as current PL estimator does, then it will result in inflated PL.
I’m sure there are more similar combinations and those might be one of causes why WLv2 is delayed. But I personally am not much worried, AI FTP does not seem to be affected by this.
1 Like
This really isn’t a great example because errant PL’s for custom workers are a known problem. Sometimes we can get good info, but often it is straight up wrong. I’d consider this an example of a bad one, and not necessarily an indictment of what WLV2 might provide with these cases.
I just want WLV2 working for TR workouts before World Wide Disaster Day 2024.
3 Likes
- In what way(s) do you hope that TR will consider this data & what changes to your plan would you want as a result?
1 Like
@mcneese.chad I made the assumption when I started using TrainerRoad that PL simply looked at total time in each respective “zone” for unstructured rides and calculated accordingly. I was always suprised that my “endurance” level stayed so low!!!
I began breaking my outdoor rides into intervals via TrainingPeaks and importing like @svens But as well as being incredibly time consuming, the results were simply not helpful.
Ideally, much like Garmins implementation, and RLGL, TrainerRoad would use the available data to provide the most suitable future workouts based on past efforts (including outside/unstructured)
But what I think most people want is for TrainerRoad to Quantify their outside effort!
A “virtual reward/slap on the back “ gamification - insights into what gains could be impacted by respective effort (ensuring proper recovery)
Progression Levels are their own worst enemy for outside/unstructured rides as they don’t reward users for their effort. Or provide a sense of advancement.
You use the term “gamification” that I dislike and don’t think is all that common really. My take is that most people want what you want: for TR to review the work done (and presumably fitness associated with those efforts in a similar way to how they use inside & outside workout data) to refine training plans via pending workout prescriptions and RLGL status over time.
PL’s or not, the goal seems to overlap more than differ from my reading. Apply a level of review of real work done (or not) and then adjust training moving forward. Simple in concept, harder in reality (usually the way despite our hopes and I gather TR’s since they didn’t expect this to take years is my guess…).
2 Likes
Agreed. And I still think that this is not so crazy hard to do. I think using 7 zones for progression levels is a big part of their issue. As well as trying to split hairs so precisely with 0.1 resolution on them. It’s not easy to try to translate essentially random input to that system.
If they had stuck to fewer levels and less precision, it would be easier to assign PLs.