TrainerRoad-like Segment analysis, annotating, saving

I’m currently opening all my TD workouts in TR, afterward, just to use their segment / interval (?) analysis feature, which I find super-helpful and fun as a beginner trying to dial in my zones.

It’s also missing a few things. For me, better implementation could add a few simple fields like:

  • RPE per segment, using Borg 0-10 scale, not TR’s bit-too-simple 1-5 scale (or a setting/pref for which scale?)
  • Comment field
  • Ability to export segments like workouts for e.g. coach / forum comments of just a given region in question. Helps get better advice, faster, if coach / peer doesn’t have to parse which part of workout to comment on. (If they need context, rider could upload both full workout and just the segment in question, too…).

Probably not a quick, easy build, but I’ve messed around with intervals.icu, and haven’t figured out how to do the same thing, there, yet, if even possible (it seems like it is, but TR is waay more instantly intuitive / usable, for me).

This feature is the only reason for me to consider a TR subscription (in theory, can’t afford right now), in addition to TD, but I’d still want the above mods.

Conversely, if this is not too hard to do in intervals.icu, or the other apps that always come up (TrainingPeaks, Golden Cheetah?), I’ve been looking for a YouTube tutorial on this, but haven’t found any.

In TR, it’s so simple and ‘just works’

Hey @Alex,

Just checking in about this. Hope this post wasn’t rude, or maybe it’s a little hard to decipher.

I’m curious if you might consider implementing this feature to your website app post-ride analysis page, at some point.

As a beginner, I use this feature a lot in TR, to look at eg heartrate drift and other averages and max values (power, cadence) from eg the first half to the second half of my zone 2 intervals.

Also, my thoughts, above, about what could be better in the TR version could be used by TrainerDay to kinda leapfrog them, and add another compelling way to differentiate TrainerDay from them.

It might be fussy to implement their “drag region” approach, but even simple “start time” and “end time” fields for creating custom post-ride analysis intervals could work great? (Similar to how your workout creator interface is way simpler than theirs but also is arguably both more advanced and easier to use?)

It’s also occurred to me that I can just design Z2 workouts where I split the Z2 interval into 2 halves, and then I would get this analysis automatically, but sometimes I drift out of parameters for a minute because I’m not glued to the stats while riding, and it’s nice to be able to choose “clean” representative regions, after the fact.

Yeah.

Thanks for any thoughts or feedback you may have about this idea.

Oh no you were not rude at all. I read it and forgot to respond.

We want to add RPE for sure… I need to see how TR does it. RPE per segment is not something I was aware of, I don’t understand this. Generally they build incredibly intuitive and well thought out “simple” training. This is the direction we want to go as well.

Overall collecting user feedback (comments/rpe) is the direction we need to move, because this is critical for any kind of adaptive training in the future.

This is likely to take a year or so before we get it done.

1 Like

TR does RPE per workout only, and uses a 5-level RPE scale.

I would love the option to do RPE per interval (including custom intervals I create post-ride) as it helps me get really granular about, eg watts and heartrate, when I’m trying to find the right “top of zone 2” gearing with my dumb trainer.

for instance, yesterday my legs felt just “edge of burn” for the first half of my Z2 interval, but then started to feel more like “slow cumulative burn” and my heart rate drifted up 10% in the second half of my Z 2 interval—so that helped me realize I may need to lower my FTP, slightly, and zone two power target on the trainer.

In addition to RPE, the ability to add comments to rides, intervals, and post hoc intervals helps me further assess if I really need to mess with my FTP and zones or if I just need to give it another week because I’m e.g. recovering from a cold or something (which I was).

These rest/recovery factors could become structured data for AI analysis, too, but I get that that might conflict with the “simple” mandate (which I totally agree with—but tradeoffs…)

So, again, with TR’s implementation, for a zone 2 interval that tips into zone 3 in the second half, it kinda sucks that my only choices are rating the whole ride as moderate or hard. That’s not quite granular enough. Even if I could rate intervals (and post hoc intervals) separately, choosing between moderate and hard is still not granular enough IMO.

Yes, Borg’s 0-10 RPE scale is less simple, but it’s maybe “as simple as is practical” for reliable polarized training. It’s still super-thoughtful and clear, IMO.

Apologies for length. I’m super passionate about this, right now!

Yes, I completely understand the problem, I’ve always struggled with the concept of RPE because at different times of the workout the RPE feels very different and how do you average them? It’s generally, you know the RPE of the hardest interval or the hardest part of the training, but still that doesn’t completely capture it either.

So I’m trying to still understand exactly what you’re saying. I’m not sure that we would do this RPE for intervals but understanding is the first step. Customers frequently bring me the best ideas.

So are you saying that you would actually track RPE for one specific interval or for multiple intervals within the ride and how would you possibly remember the RPE for multiple intervals after the ride was done?

I really love that you’re setting your FTP based on your zone two results super cool. I guess I’m partially seeing that the way that you’re stating it in written format in comments, AI could actually figure out where to apply that to a specific interval in the future meaning you don’t necessarily need to know Exactly where or you could specify within your comments just some time sections you know at 37 minutes and 27 seconds I felt like this… Again, just brainstorming

1 Like

Re Zone 2– thanks!

Re AI interpolation of comments: Yes, that could work. But the human visual analysis aspect is even more cool, IMO.

I encourage you to play around with TR’s implementation of post-ride custom intervals to better understand the full value / use cases.

The way you can quickly zoom in on the graphs of multiple post-hoc segments, and also see the summarized stats for these segments, instantly—this is also really valuable and fun.

Ok I will check it out next time I subscribe :slight_smile: Unless you want to create me a video, which would even be cooler from my perspective…

1 Like

Re making a video: Yes, happy to do that at some point, but might take a while.

In the meantime, though, here are two very quick resources demoing the feature–first video is very similar to how I use it:
https://www.youtube.com/watch?v=ryRM8P2mjmw (~1 min long)

https://support.trainerroad.com/hc/en-us/articles/360002338952-How-to-Analyze-Ride-Data (scroll down or click in outline to ‘create and analyze custom intervals’)

Thanks for sharing that is perfect. This analysis is a beautiful feature the way they have done it. More attractive than anyone else’s implementation. Until we get bigger we are avoiding any real analysis as it gets expensive and our goal is to stay low cost.

But I see three core things in your request

  1. RPE + Comments
  2. Analysis
  3. RPE/Analysis Sharing

The analysis part is done by so many apps that have so many deep features around this. I know it is nice that TR does it in one nice simple app but they are $22 / month… We have to look at the most core/basic elements and simplicity and try to offer that.

Just to make sure we are on the same page. We do have a simplified version of that on our website. see below

Strava premium, intervalsICU, TrainingPeaks, GoldenCheetah all have great in-depth analysis.

So in my mind the part I am most interested in at this point is the RPE and comments. Meaning this is critical end of ride data that should be captured, and could help drive AI or even the individual and likely some additional simple reporting on this data.