Modeled FTP - Feature Request

Hi TR!

Loving the new calendar and forums - well done!

I wanted to suggest that as a next step, TR implements a Modeled FTP. This would estimate (and update) FTP based on values extracted from MMP power duration curve, which is a feature already present in the software.

The smarts for this are pretty well known (it’s been done in GoldenCheetah, WKO4+ and also on the newest Stages head units).

This could ultimately eliminate the need for any kind of formal testing, once a rider has enough of their history in the system.

It would be important that mFTP is blanked if there is insufficient recent history (or ride data) to establish a quality metric

As a side note - any reason we couldn’t get a Fitness/Fatigue/Form graph too? Doesn’t need to be called CTL/ATL/TSB - many sites simply call it LTS (Long Term Stress)/STS (Short Term Stress)/SB (Stress Balance).


1 Like

This doesn’t seem like it would work to me. All workouts are based on your current FTP, so no section should push you above that threshold far or long enough to establish a new FTP during any actual workout. Unless I’m just misunderstanding what you’re suggesting, I don’t see how it’s possible because of how the program is laid out.


Just guessing here, but it could maybe work if they’re analysing race efforts…

@Summerson and @sjooons: Modeled FTP isn’t theoretical, it’s been implemented as part of several other platforms. The key of course is that it is modeled and this is based on mean maximal power at varying lengths of time (hence needing to have enough of a history of varied riding lengths/intensities for the model to have enough good data).

The models out there are producing mFTP results within one or two watts of results of more formal FTP testing protocols. Which is well within the margin of error you’d expect by doing the tests themselves.

1 Like

I’m not sure this would add much value to the platform. The current modelled thresholds in other platforms look at a all rides and these are not part of a ‘training’ platform. TrainerRoad is a service that provides workouts for you to complete and improve your perforamance. These workouts alone would never create a new ftp; unless you are doing a formal test.
Of course you could process the external rides and look at these for increased power but in my opinion these should only notify of a possible change in ftp.
One thing to remember with all of these modelled metrics is that you need to feed the model. So if you are head deep into a winter dose of endurance miles then a modelled ftp would drop. Which in turn would drop your workouts, dropping your modelled ftp further.


All good points @Adam.

mFTP wouldn’t replace an FTP value obtained via a formal test. It would simply provide another data point based on real world riding over (relatively recent) period of time.

Agree that if all one is doing is winter base miles, the model doesn’t get the inputs it needs and of course the standard testing protocols should be used.

But wouldn’t it be great if, after a Summer of riding (indoors and outdoors) you get a little popup that says "hey, we think your FTP has increased based on your riding - do you want to accept this?

Also, I think TrainerRoad is no longer simply “a service that provides workouts for you to complete and improve your perforamance”. By incorporating outdoor rides, providing seasons to compare performance trends and now lately this calendar it’s clear (and I’m loving) that TR is evolving more into a portal for planning, recording, managing and analysing performance in general.

Before the ramp test I’d have agreed…However the ramp test is so much faster and less impactful on training then happy to do it more regularly to keep an up to date ftp

1 Like

A big +1 on adding the fitness&freshness data to the graph. I would really like that!


Another +1 for LTS, STS and SB curves…