Strava's training load is based on "Bike Score" as opposed to TSS. Bike Score is a function of xPower which is similar to Normalized Power but is based on a 25 second weighed exponential moving average. These are all Phil Skiba's metrics, which are similar to Coggan/WKO's but slightly different. If you search, you can find Skiba's articles where he explains the physiological motivations for his calculations.

My understanding is that TSS, NP, etc., are copyrights of Coggan's or Training Peaks, and can't be freely used in a commercial product like Strava.

Strava Fitness and Freshness chart is not a serious tool, at least not the last time I looked into it, because your historical Bike Score, short term stress, and long term stress are all recalculated every time you update your FTP.

 

Nice explanation, bloodninja. I think it is really dumb that Strava updates the entire Fitness and Freshness chart every time I update my FTP. I mean, come on...just because I improved this month, it doesn't mean I was that strong a few months ago! Geez!

 

I know this is an old thread but what if you used strava's F&F tool exclusively?
I agree that not being able to look at your TSS suuucccks. But I understand Strava's reasoning why...and that suuucccks too. You think at this point since Coogen/TP's TSS and other "patneted" metrics have been out for so long that they'd just let it be acceptable to used as a 3rd party thing, opr how about pay for it as a premium user? I don't know the real specifics but, anyway.

If you talk to any coach that responds to whiny people taling about power numbers varrying from devise to devise and small varients, most will tell you consistantly...as long as you are calibrating often, just use the same meter and cosistantly monitor those numbers.
Why can't you do the same with Strava's F&F?

 

Last I knew, the strava stuff changed each time you adjusted your ftp. So as your season progresses and you make gains (either over the course of the season or season over season) it applies that ftp to all of your prior rides which changes the preceding "tss" values.

I'm not savvy enough to know if that's a more correct way to look at it (allegedly the formula for CTL uses the past ~40days maybe 6 weeks) so that difference in calculation probably doesn't lead to a ton of difference but...????

Maybe that's changed?

 

Last I knew, the strava stuff changed each time you adjusted your ftp. So as your season progresses and you make gains (either over the course of the season or season over season) it applies that ftp to all of your prior rides which changes the preceding "tss" values.

I'm not savvy enough to know if that's a more correct way to look at it (allegedly the formula for CTL uses the past ~40days maybe 6 weeks) so that difference in calculation probably doesn't lead to a ton of difference but...????

Maybe that's changed?

IMO this is an incorrect way to quantify TSS. TSS assesses current effort against current zones to quantify current training stimulus. If you insert fictional numbers for 'current zones' (and tomorrow's numbers are fictional numbers as far as I'm concerned) you'll have a fictional TSS as a result.

I use TrainingPeaks for my serious training analysis work, sometimes GoldenCheetah. But it's fun to see the Suffer Score on Strava at times. But that's all I use it for: fun.

 

As a free alternative you can use goldencheetah to use the PMC. It can basically track it the same way as training peaks do, only drawback is you got to customize it a bit and its not intuitive, but there are many guides available for it.

 

This is fascinating.

What I have found is, Strava does not store a history of FTP and the TSS is calculated at time of display based on Normalized Power and Elapsed Time. This is why adjusting your FTP changes the F&F chart.

I was thinking about this stuff over the past week because a Strava connected app I am developing calculates F&F using the same formula a Strava. I was trying to decide, do I calculate on current FTP or historical FTP (the TSS you had when you did the ride).

On the one hand, people might think mine is broken if it does not match Strava, on the other hand, I do think it should be based on when the activity was done and not recalculated on current FTP.

 

It's weird how Strava doesn't take seriously in this regard. You would think they would at least try to get things right, but no, they simply don't care about skewed results. They just care about the social aspect of the sport.

So use it for fun and to brag about your numbers, but not for proper training.

 

I think Walt is on the right track here. Adding to that, none of this can be truly accurate, because no system takes into account external stresses that can affect stress and recovery, etc.

I think numbers are relative to the system and can work as long as you stick to that system. I don't think that makes anyone less serious, it just means they are using different tools.

I know based on my software that Strava is using the same math as TrainingPeaks published (no clue if it is what they actually use internally) because I used TP's formula for TSS and it returns the same TSS as Strava. Also, using the TP formula for F&F, I can use 1000 activities based on a single FTP and get the exact same F&F numbers as the Strava chart. Considering each F&F number is calculated on a six weeks average of the previous F&F numbers, my results are not an accident.

Why Gramin comes up with a different number, makes me wonder what formula they are using. They say it is based on ET, FTP, and NP, but maybe they are doing something different there. Also, Auto-pause or Smart recording could mess things up if they are not accounted for.

 

I think Walt is on the right track here. Adding to that, none of this can be truly accurate, because no system takes into account external stresses that can affect stress and recovery, etc.

I think numbers are relative to the system and can work as long as you stick to that system. I don't think that makes anyone less serious, it just means they are using different tools.

I know based on my software that Strava is using the same math as TrainingPeaks published (no clue if it is what they actually use internally) because I used TP's formula for TSS and it returns the same TSS as Strava. Also, using the TP formula for F&F, I can use 1000 activities based on a single FTP and get the exact same F&F numbers as the Strava chart. Considering each F&F number is calculated on a six weeks average of the previous F&F numbers, my results are not an accident.

Why Gramin comes up with a different number, makes me wonder what formula they are using. They say it is based on ET, FTP, and NP, but maybe they are doing something different there. Also, Auto-pause or Smart recording could mess things up if they are not accounted for.

Again IMO, FTP is just one of many markers in the sand. Most of my work is less about FTP per se and more about increasing wattage on sub-5min efforts, or learning how to suffer through long tempo efforts then accelerating etc., or something else I've found as a weakness. Someone on the XC racing thread made the comment that '...we don't just mail our FTPs in to race organizers', which is spot on. But what I'd have a problem with in the model expressed above, where FTP (or critical power figures at any time break) are retrospectively applied to prior efforts as your engine grows, is that (a) you lose any sensible assessment of CTL and related metrics (form, fatigue if you're on the PMC chart) and (b) you lose the ability to benchmark efforts. For me, pulling off a 5 hour ride at 80%+ of FTP is a real win no matter what the FTP is. But if I get stronger over time and FTP is applied retrospectively, then what used to be a 80% FTP 5-hour ride could drop to 70% or less, and that doesn't tell the right story. And I use stuff like that to get motivated, like 'oh yeah, I've done this before and it was tough but I can definitely make it through'.

Strava isn't going to listen to us anyway, but it's fun to overanalyze some of these things.

 

What about using the stravistics add-on TRIMP score's? My info is only HR based as well since I do not own a PM.

 

Have no experience with the TRIMP add-on.

 

Stravistics is a free chrome extension for strava that provides more detailed data including TRIMP score. It can work with a HRM or PM. I am interested if anyone else has any experience with it. It’s updated often and seems like there has been a lot of work put into it.