Jump to content
BogFrog

Training Peaks, ATL, CTL,TSB & ABC

Recommended Posts

I use the stryd.  I find it over values your power / TSS

For example a 10min run off the bike easy I got a 25 TSS today.

a 60 minute easy bike with a power meter (Stages) I got a 28 TSS (Yes very easy ride with a mate back from a knee injury.)


I did an hour swim on Monday night for a TSS of 96. 

Edited by Peter

Share this post


Link to post
Share on other sites
1 hour ago, Peter said:

I use the stryd.  I find it over values your power / TSS

For example a 10min run off the bike easy I got a 25 TSS today.

a 60 minute easy bike with a power meter (Stages) I got a 28 TSS (Yes very easy ride with a mate back from a knee injury.)


I did an hour swim on Monday night for a TSS of 96. 

How are you setting threshold power value with the Stryd?

Share this post


Link to post
Share on other sites
57 minutes ago, Alex Simmons said:

How are you setting threshold power value with the Stryd?

On what it suggests.

Share this post


Link to post
Share on other sites
56 minutes ago, Peter said:

On what it suggests.

Is that the same as the power it reads when running at your threshold pace (e.g. maximal you can maintain for ~40-60 minutes)?

It should be different to your cycling power, typically higher.

Share this post


Link to post
Share on other sites

I've been using the Stryd as well and find it typically is very low when compared to cycling numbers. Understandable given that cycling's TSS and Stryd's RSS were not made to be measured against each other however confusing if you aren't aware considering they 'appear' to measure the same thing. 

I've started comparing RSS and then taking the power data and running it as if it were a cycle and how TSS metrics would come out. Given the exact same threshold value it seems to be suggesting that the run causes a lower stress on the body than an equivalent ride. 

As an example, I raced Sydney marathon last year. With a threshold of 325W, Stryd gives me a RSS of 205. I ran the same file as if it were a bike ride and it results in a TSS of 237. 

 

This is arguably the biggest issue with metrics like this is that they all 'seem' to measure the same thing but were never designed in conjunction with each other so with what is currently available, they simply can't be combined to form a single number to provide any sort of usable information.

  • Like 1

Share this post


Link to post
Share on other sites

Yes. 

Numbers must be looked at in conjunction with RPE. And not only power HR/Power comparisons are valid. 

Share this post


Link to post
Share on other sites
22 hours ago, AA7 said:

Settle down Cranky, you'll be fine. I heard stressing about decreasing fitness actually makes fitness decrease even more.

Ahhhhhhhhhhh 😣

21 hours ago, skinnee said:

You should set some heart rate setting for your strength sessions if you're  OCD like me , it will add a stress score to your overall fitness, though will be minor.

 

And yep I have a had a foot injury for a few months now and my bike and run looks like that. 

Thanks.  Didn't know you could do that. It's on my to do list

 

What did you do to your foot? 

17 hours ago, Alex Simmons said:

How are you setting threshold power value with the Stryd?

What's stryd?

Share this post


Link to post
Share on other sites
38 minutes ago, Cranky said:

Ahhhhhhhhhhh 😣

Thanks.  Didn't know you could do that. It's on my to do list

 

What did you do to your foot? 

What's stryd?

Running power meter... and seen you can't run you don't need one...

Share this post


Link to post
Share on other sites
21 minutes ago, pieman said:

Running power meter... and seen you can't run you don't need one...

But I sense another thread asking if she should get one when she can run again :)

  • Haha 1

Share this post


Link to post
Share on other sites
8 minutes ago, AA7 said:

But I sense another thread asking if she should get one when she can run again :)

Yep well this could be the edge she needs on the run to take you out... 😁

Share this post


Link to post
Share on other sites
13 hours ago, BJ7 said:

As an example, I raced Sydney marathon last year. With a threshold of 325W, Stryd gives me a RSS of 205. I ran the same file as if it were a bike ride and it results in a TSS of 237. 

That suggests either the inputs are not the same (absolute or relative), the algorithm used is not the same, or a bit of both.

In any case, your concerns about how to use the data is warranted and makes sorting out what is/isn't useful or how to make it useful a little more challenging.

With any tool or data analysis we are seeking additional insight (over and above what we can reasonable/effectively/practically ascertain via existing tools) and actionable intelligence (IOW is what's measured and reported informing you about what if any changes to your training or racing strategy you should make that you would not have determined already via existing means?).

Keep in mind that training loads are only one element of the data.

Share this post


Link to post
Share on other sites
18 minutes ago, Alex Simmons said:

That suggests either the inputs are not the same (absolute or relative), the algorithm used is not the same, or a bit of both.

In any case, your concerns about how to use the data is warranted and makes sorting out what is/isn't useful or how to make it useful a little more challenging.

With any tool or data analysis we are seeking additional insight (over and above what we can reasonable/effectively/practically ascertain via existing tools) and actionable intelligence (IOW is what's measured and reported informing you about what if any changes to your training or racing strategy you should make that you would not have determined already via existing means?).

Keep in mind that training loads are only one element of the data.

That was my initial thoughts however believe i have got it right. I have calculated using both Average power and NP to check the differences as well. Threshold in all methods is set at 325W. The power data is all from the same file so no changes there. Duration is obviously constant. 

TSS formula I have calculated using 3 methods (Training Peaks, Today's Plan and manual) and they all confirm each other. 

RSS was calculated by just plugging the file into Stryd itself. 

Also this wasn't an isolated discrepancy and occurs on every run.

Share this post


Link to post
Share on other sites

Then the Stryd algorithm isn't the same.

I just looked it up and can immediately see the problem.

Quote

RSS = 100 x training duration x (Power/CP)^K

Critical power (CP): your maximal average sustainable power over one hour of running, also referred to as “lactate threshold” power; this is used as your performance baseline.

https://blog.stryd.com/2017/01/28/running-stress-score/

Firstly they don't use Normalized Power in their calculation, instead they just use Power. This is a major flaw as the use of NP is fundamental to the calculation of Intensity Factor and hence stress score.

Using Power instead of NP means no attempt is made to account for changes in intensity during a run nor consider the time course of nor the curvilinear nature of the metabolic and physiological responses and stresses involved with exercise at different intensity levels. This would become immediately apparent to anyone who does any form of interval training. Steady state 200W for 30-min <> 10 intervals comprising 1-min at 400W followed by 1-min rest. 

As we know AP is at most equal to NP and typically somewhat less than NP especially as power output becomes more variable, so the Stryd algorithm will provide a lower stress score than if it used a Normalized Power value.

 

Secondly, Stryd then apply a different and mystery exponent "K" to their bastardised Intensity Factor (Power/CP), which is not revealed but they do say the exponent is significantly higher than that used to calculate TSS.

It would be possible to back calculate what they do use from a power file, however there are consequences to using a higher exponent and for all sub-threshold runs it will suppress the RSS value more than the TSS calculation does.

This is completely counter to what they claim:

Quote

Indeed, the coefficient K for running RSS is significantly higher than that of cycling TSS, reflecting the higher stressing effect of intensive run workouts.

e.g. in TSS, the IF is squared.

Say IF = 0.8

For one hour TSS = 100 x (0.8)^2 = 64

If we use a higher exponent, say 3 then

For one hour RSS = 100 x (0.8)^3 = 51

The only time their formula results in a higher stress value than TSS is when the Power is higher than CP for the entire run.

 

So not only are they not accounting for the impact of intensity variability during a run, which suppresses their calculation of RSS, they are in fact also suppressing RSS values compared with TSS for all runs conducted at any sub threshold power/pace.

 

Their definition is CP is wrong as well, but that's somewhat less of a crime in the context of explaining differences between their RSS and TSS.

 

I don't know who at Stryd came up with that, but their attempt to come up with an RSS that reflects running has little basis is physiology, and the maths they use does not achieve what they claim it does.

It's no wonder everyone is confused by it.

  • Like 1

Share this post


Link to post
Share on other sites
13 minutes ago, Alex Simmons said:

Then the Stryd algorithm isn't the same.

I just looked it up and can immediately see the problem.

https://blog.stryd.com/2017/01/28/running-stress-score/

Firstly they don't use Normalized Power in their calculation, instead they just use Power. This is a major flaw as the use of NP is fundamental to the calculation of Intensity Factor and hence stress score.

Using Power instead of NP means no attempt is made to account for changes in intensity during a run nor consider the time course of nor the curvilinear nature of the metabolic and physiological responses and stresses involved with exercise at different intensity levels. This would become immediately apparent to anyone who does any form of interval training. Steady state 200W for 30-min <> 10 intervals comprising 1-min at 400W followed by 1-min rest. 

As we know AP is at most equal to NP and typically somewhat less than NP especially as power output becomes more variable, so the Stryd algorithm will provide a lower stress score than if it used a Normalized Power value.

 

Secondly, Stryd then apply a different and mystery exponent "K" to their bastardised Intensity Factor (Power/CP), which is not revealed but they do say the exponent is significantly higher than that used to calculate TSS.

It would be possible to back calculate what they do use from a power file, however there are consequences to using a higher exponent and for all sub-threshold runs it will suppress the RSS value more than the TSS calculation does.

This is completely counter to what they claim:

e.g. in TSS, the IF is squared.

Say IF = 0.8

For one hour TSS = 100 x (0.8)^2 = 64

If we use a higher exponent, say 3 then

For one hour RSS = 100 x (0.8)^3 = 51

The only time their formula results in a higher stress value than TSS is when the Power is higher than CP for the entire run.

 

So not only are they not accounting for the impact of intensity variability during a run, which suppresses their calculation of RSS, they are in fact also suppressing RSS values compared with TSS for all runs conducted at any sub threshold power/pace.

 

Their definition is CP is wrong as well, but that's somewhat less of a crime in the context of explaining differences between their RSS and TSS.

 

I don't know who at Stryd came up with that, but their attempt to come up with an RSS that reflects running has little basis is physiology, and the maths they use does not achieve what they claim it does.

It's no wonder everyone is confused by it.

 

Exactly correct and I worked back to the same conclusion that by increasing the exponent, they're actually achieving the opposite result of what they claim to be achieving. What I am surprised about though is how they managed to get that wrong and that it's not been found by anyone? 

I know on a Golden Cheetah forum someone attempted to reverse engineer the value for "K" but don't recall off the top of my head what it resulted in. 

 

Share this post


Link to post
Share on other sites

So basically it;'s a cadence and step counter and the power side is useless?

 

Share this post


Link to post
Share on other sites
6 hours ago, pieman said:

Yep well this could be the edge she needs on the run to take you out... 😁

Well, maybe I'll get one too.

My run times are on the improve again. Humble brag here, I don't do it much. Started doing parkrun in July last year, 22.30 first up, 20.26 last week. 

  • Like 2

Share this post


Link to post
Share on other sites
1 hour ago, Peter said:

So basically it;'s a cadence and step counter and the power side is useless?

 

 

No I don't think so - I'm quite liking the information its giving me and how I can use it. Just the training load stuff I don't think holds much weight. 

Share this post


Link to post
Share on other sites
3 hours ago, Peter said:

So basically it;'s a cadence and step counter and the power side is useless?

What I'm referring to is issues with the post hoc analysis of the Stryd data, not the data itself. In this case Stryd's calculation of RSS seems somewhat flawed to me and runs counter to their claims.

I suspect there are a range of insights to be gained from thoughtful analysis of the Stryd power data. The greater complexity involved in assessment of running power means it may take a little longer to sort out the wheat from the chaff.

Share this post


Link to post
Share on other sites
5 hours ago, BJ7 said:

Exactly correct and I worked back to the same conclusion that by increasing the exponent, they're actually achieving the opposite result of what they claim to be achieving. What I am surprised about though is how they managed to get that wrong and that it's not been found by anyone?  

Well we know at least two people have :)

Seriously though, it took me about 3.5 seconds to spot it when I saw the formula and their accompanying notes.

Keep in mind that because they are not using Normalized Power or a reasonable proxy, the "Intensity Factor" value being raised to that exponent will also be lower than it would be with a TSS formula.

It's a double whammy suppression of all sub threshold RSS values when compared with the TSS formula.

It might be a triple whammy if they calculate CP inappropriately as well, and given their incorrect definition well I'm not laying any money on that one either.

5 hours ago, BJ7 said:

I know on a Golden Cheetah forum someone attempted to reverse engineer the value for "K" but don't recall off the top of my head what it resulted in. 

I just searched the GC forum and they came up with an exponent of K = ~3.13 however based on Stryd's own RSS/hour and intensity values it's not a perfect fit (R^2 of 0.98). It might be due to some rounding of the values in the table than K being a customised variable as opposed to a constant.

I can feel a Wattage forum post coming on...

Share this post


Link to post
Share on other sites

I posted to Wattage forum and am learning a bit more, as I know there are some  smart people who will both enlighten and correct me where my understanding is lacking.

Ale Martinez pointed out he attempted to back calculate K but used an exponential function to simulate it to fit a curve, and not a power function, hence his estimated value for K is not the exponent Stryd are using.

I'd have thought it wouldn't be all that hard to work out given values for Power, CP, RSS and duration.

Share this post


Link to post
Share on other sites
On 08/02/2018 at 5:47 AM, pieman said:

Running power meter... and seen you can't run you don't need one...

Lol. Gee thanks! 😭

23 hours ago, AA7 said:

But I sense another thread asking if she should get one when she can run again :)

Lol. I have no money left! 

23 hours ago, pieman said:

Yep well this could be the edge she needs on the run to take you out... 😁

Oooo. If that's the case .... 😄

16 hours ago, AA7 said:

Well, maybe I'll get one too.

My run times are on the improve again. Humble brag here, I don't do it much. Started doing parkrun in July last year, 22.30 first up, 20.26 last week. 

Shit! 

Share this post


Link to post
Share on other sites
On 2/1/2018 at 11:00 PM, skinnee said:

TSB is training stress balance and is the difference between the two and is an overall indicator of the training stress you are currently under , this number can help fine tune your training.

Hi - Still learning this stuff, as I leave it to the coach, however the above statement appears to be a little incorrect - not sure what affects it however my current state is ATL - 187, CTL - 148 which should leave the TSB as being -39, however my TSB is actually -55, that seems to be quite a variance.

image.png.b18da70d34a2bd9f4a4e9b9d12dbcaf8.png

thanks Stephen

Share this post


Link to post
Share on other sites

Create an account or sign in to comment

You need to be a member in order to leave a comment

Create an account

Sign up for a new account in our community. It's easy!

Register a new account

Sign in

Already have an account? Sign in here.

Sign In Now

×