Performance Manager Question for Triathletes

TMD said:

I was curious if and/or how triathletes are using the new Performance Manager chart in CP WKO+? More specifically, I am curious how different people are incorporating running workouts? Do you assign an IF based on heart rate for running workout and then use the IF and time to calculate TSS? or is the calculation of TSS based on a percieved exertion? or do you not attempt to quantify the stress of running in the Performance Manager chart? Intuitively I would think that you would want to capture some stress value if you include running in your weekly workouts?

My intial thought is to create two Performance Manager charts: 1) Would be based on just cycling with downloads from my PT SL 2) Create a second chart that attempts to quantify the stress of running by assigning an intensity factor and then manually calculating TSS. Over time, I could comapre the two different charts with my actual results and decide which has been more accurate in tracking fitness peaks - i.e. it would be a lot of science and some art.

As a bit of background, I do a mix of road racing and triathlons. Road racing in the spring, and then I switch gears to triathlons in the late summer/fall.

I think some have began mixing Cycling TSS, with running GOVSS. That being said, I don't know what GOVSS is, and got only very few hits on google about it.

I might even have the wrong spelling. I donno, I read Bob Tobin mentionning something about it on a tri-forum. I'd expect it to be related to Running speed rather than power of course, and I'd also expect it to make use of Forerunners products.

Not sure there's any solution out there for swimming though (other than trying to come up with a TRIMP base Power Manager system).

I would love this Performance Manager project to used with TRIMP. I hope it's going to happen someday 🙄

TMD,

I just finished reading this here .. http://www.cyclingpeakssoftware.com/power411/performancemanagerscience.asp

It can be used with TRIMP so triathlon can make use of Performance manager. I haven't installed latest CyclingPeak yet, but I remember having read that it's possible to input running workouts in it.

Maybe it can process a PM chart out of HR data maybe?

SolarEnergy said:

TMD,

I just finished reading this here .. cyclingpeakssoftware.comperformancemanagerscience.aspOpen ↗

It can be used with TRIMP so triathlon can make use of Performance manager. I haven't installed latest CyclingPeak yet, but I remember having read that it's possible to input running workouts in it.

Maybe it can process a PM chart out of HR data maybe?

I admittedly do not know a lot about TRIMP, but an article that I pulled up during a Google search last night outlined the calculation as follows:
(Avg HR - Resting HR) / (Max HR - Resting HR) = Relative Intensity
Relative Intensity * time = TRIMP

If I'm attempting to come up with a proxy for TSS for running workouts to manually input into the Perfomance Manager chart, could I assume that the Relative Intensity in the equation above is roughly equivalent to the IF calculated by CP WKO+ from a powermeter download?

TMD said:

I admittedly do not know a lot about TRIMP, but an article that I pulled up during a Google search last night outlined the calculation as follows:
(Avg HR - Resting HR) / (Max HR - Resting HR) = Relative Intensity
Relative Intensity * time = TRIMP

If I'm attempting to come up with a proxy for TSS for running workouts to manually input into the Perfomance Manager chart, could I assume that the Relative Intensity in the equation above is roughly equivalent to the IF calculated by CP WKO+ from a powermeter download?

No it can't be.

In order to end up with an IF of 1, all you need to do is ride at 100% FT

In order to end up with a RI of 1, what you'd have to do is ride at 100% of VO2Max, if I follow your calculation.

If we simplify things as much as we can (not sure that we will keep decent accuracy by doing so), we could probably calculate relative intensity with the following equation instead
Avg_HR / HR at FTP 🙄

But I am not sure what we'd do with this value. If I am not mistaken, wko+ can't accept TRIMP data as input for the Powermanager function at the moment. So any attempt to use TRIMP would have to take place outside CyclingPeaks (at least for now).

And so once outside CP context, I am not sure of the relevance of Relative Intensity variable. Take for instance TRIMP calculated by zone, some sort of 'relative intensity' is already weighted in this calculation.

TMD said:

I admittedly do not know a lot about TRIMP, but an article that I pulled up during a Google search last night outlined the calculation as follows:
(Avg HR - Resting HR) / (Max HR - Resting HR) = Relative Intensity
Relative Intensity * time = TRIMP

If I'm attempting to come up with a proxy for TSS for running workouts to manually input into the Perfomance Manager chart, could I assume that the Relative Intensity in the equation above is roughly equivalent to the IF calculated by CP WKO+ from a powermeter download?

If you want it to be similar to TSS, you'll need to divide the relative intensity of your workout by your threshold relative intensity to get an IF.

Then IF is squared and multiplied by time * 100.

SolarEnergy said:

No it can't be.

In order to end up with an IF of 1, all you need to do is ride at 100% FT

In order to end up with a RI of 1, what you'd have to do is ride at 100% of VO2Max, if I follow your calculation.

If we simplify things as much as we can (not sure that we will keep decent accuracy by doing so), we could probably calculate relative intensity with the following equation instead
Avg_HR / HR at FTP 🙄

But I am not sure what we'd do with this value. If I am not mistaken, wko+ can't accept TRIMP data as input for the Powermanager function at the moment. So any attempt to use TRIMP would have to take place outside CyclingPeaks (at least for now).

And so once outside CP context, I am not sure of the relevance of Relative Intensity variable. Take for instance TRIMP calculated by zone, some sort of 'relative intensity' is already weighted in this calculation.

Thanks for your responses. Is there some rough equivalent to FTP based on heart rate? For example, could I assume that my FTP is roughly equivalent to 80% (or some other percentage) of my max HR during a running workout? If that is the case, and I completed a running workout exactly at 80% of my max HR, then the IF would be 1.0

As you pointed out, the problem with the orginal equation would be the inability to complete a workout with an intensity factor of 1.0. I was thinking, however, that while the above equation would be conservative (or maybe flawed) it would still allow one to capture some kind of stress function from running workouts that would otherwise be ignored if one is only loading in Cycling Power Numbers into the Performance Manager.

It seems to me that you could do pretty well for running by just using a normalized version of the kcal estimate that comes out of your heart rate monitor (as long as you set up the HRM correctly with weight, age, and max and rest hr etc). I guess the question becomes how to normalize the kcal estimate to get something equivalent to TSS. Here's an idea. Take your cycling FTP and multiply by 3.6. Call this KJ. This is approximately your calories used by cycling at FTP for an hour. Then all you do is take your HRM calorie estimate from your run, multiply by 100 and divide by KJ (formula is KCAL*100/KJ, where KCAL is from the HRM and KJ=FTP*3.6). This gives you an approximate TSS score for your run I think. It doesn't do as good a job as TSS at taking into account high variance workouts but it should be pretty good. (It doesn't adjust upward using the 4th power, for example, but neither will any calc using average hr.) Nothing based on HR alone will do as well as TSS.

lanierb said:

It seems to me that you could do pretty well for running by just using a normalized version of the kcal estimate that comes out of your heart rate monitor (as long as you set up the HRM correctly with weight, age, and max and rest hr etc). I guess the question becomes how to normalize the kcal estimate to get something equivalent to TSS. Here's an idea. Take your cycling FTP and multiply by 3.6. Call this KJ. This is approximately your calories used by cycling at FTP for an hour. Then all you do is take your HRM calorie estimate from your run, multiply by 100 and divide by KJ (formula is KCAL*100/KJ, where KCAL is from the HRM and KJ=FTP*3.6). This gives you an approximate TSS score for your run I think. It doesn't do as good a job as TSS at taking into account high variance workouts but it should be pretty good. (It doesn't adjust upward using the 4th power, for example, but neither will any calc using average hr.) Nothing based on HR alone will do as well as TSS.

kcal is proportional to total work performed. It does not account for intensity. TSS, on the other hand, is based on IF^2 and is therefore non-linearly related to work or kcal. You'd be way off on workouts that are not near threshold pace.

lanierb said:

It seems to me that you could do pretty well for running by just using a normalized version of the kcal estimate that comes out of your heart rate monitor (as long as you set up the HRM correctly with weight, age, and max and rest hr etc). I guess the question becomes how to normalize the kcal estimate to get something equivalent to TSS. Here's an idea. Take your cycling FTP and multiply by 3.6. Call this KJ. This is approximately your calories used by cycling at FTP for an hour. Then all you do is take your HRM calorie estimate from your run, multiply by 100 and divide by KJ (formula is KCAL*100/KJ, where KCAL is from the HRM and KJ=FTP*3.6). This gives you an approximate TSS score for your run I think.

I wonder if it's not this route that Dr. Skiba has taken. I think he is the one who came up with GOVSS. SS probably stands for Stress Score, but as for GOV... I donno.

Thanks for your suggestion !

** edit **

bon cimonak !

Gravity Ordered Velocity Stress Score

Finally found the data... available right here on Dr. Phil Skiba's web site.
http://www.physfarm.com/

That doesn't solve the case for swimming, but quite frankly, one shouldn't be spending too much sugar in this department anyway.

JustCurious said:

kcal is proportional to total work performed. It does not account for intensity. TSS, on the other hand, is based on IF^2 and is therefore non-linearly related to work or kcal. You'd be way off on workouts that are not near threshold pace.

First, that's why I said this:

"It doesn't do as good a job as TSS at taking into account high variance workouts but it should be pretty good. (It doesn't adjust upward using the 4th power, for example, but neither will any calc using average hr.) "

[The fourth power comes from the normalized power calculation.]

However, I'm not sure I agree that you would be "way off" (though it depends on what you mean by way off). For example, I just ran a correlation between TSS and KJ for my last year's workouts and got 0.97. Squaring that gives you an R^2 of 0.94, which means that KJ alone explains 94% of the variation in TSS. It's not perfect, as I said, but, by my definition anyway, that's not way off either. It's a pretty good first order approximation. Note that you can't do any better using average HR because that ignores the variation within the workout as well, so to do better you're going to have to do something like a "normalized hr" calculation (not unlike the normalized power calc).

lanierb said:

However, I'm not sure I agree that you would be "way off"

My initial understanding of Dr. Skiba's works, is that he does seem to use energy costs as a basis for his calculations.

The Development of the GOVSS Algorithm - By Philip Friere Skiba DO said:

"It has been determined that the energy requirement (C) to cover any given distance is essentially
independent of velocity, is consistent within individual athletes, and is equivalent to between 3.6-
4.2 J· Kg-1 · M-1. (Margaria et al 1963, Pugh 1970, Cavagna and Kaneko 1976, Fellingham et al
1978). Importantly, C also varies with the slope of the running surface (i) according to a 5th order
polynomial regression (Minetti 2002):"

But I think he does that in a way that is similar to Dr. Coggan's TSS. He records what is the running velocity every 120s (I think), and process it against the athlete's individual running velocity at LT (or whatever the name he gives to this threshold).

So you weren't that "way off"

lanierb said:

First, that's why I said this:

"It doesn't do as good a job as TSS at taking into account high variance workouts but it should be pretty good. (It doesn't adjust upward using the 4th power, for example, but neither will any calc using average hr.) "

[The fourth power comes from the normalized power calculation.]

However, I'm not sure I agree that you would be "way off" (though it depends on what you mean by way off). For example, I just ran a correlation between TSS and KJ for my last year's workouts and got 0.97. Squaring that gives you an R^2 of 0.94, which means that KJ alone explains 94% of the variation in TSS. It's not perfect, as I said, but, by my definition anyway, that's not way off either. It's a pretty good first order approximation. Note that you can't do any better using average HR because that ignores the variation within the workout as well, so to do better you're going to have to do something like a "normalized hr" calculation (not unlike the normalized power calc).

That makes a lot of sense. Thanks for your input lanierb. As I mentioned in the original post, I plan on creating two Performance Manager Charts: one with just cycling PowerTap downloads and a second that incorporates a manually calculated TSS for running workouts. I realize that there is a good deal more "art" incorporated into the running TSS, but hopefully it will result in capturing a close approximation of stress level (and it will certainly be better than assuming a zero value by ignoring running workouts). The ultimate proof will be how well my results correlate to the Performance Manager values.

If you use a Polar HRM and their PPP software, there is a facility to set up a Polar 'TRIMP' - they refer to it as 'Exertion' and is basically time spent in each HR zone * HR zone factor.

Its fairly basic but seems to work well for measuring relative intensity.

I use it for rides where I don't have a power meter fitted to my bike and then use the Exertion Count and Average HR for that ride to compare against known 'power' rides (both road and trainer) to estimate IF and NP. Seems to work quite well (against back to back testing) 😎

Couldn't see why you couldn't do the same thing for running (and swimming if you use a HRM in the pool) 😉

MY02_STi said:

If you use a Polar HRM and their PPP software, there is a facility to set up a Polar 'TRIMP' - they refer to it as 'Exertion' and is basically time spent in each HR zone * HR zone factor.

Its fairly basic but seems to work well for measuring relative intensity.

I'm using this method, for validation I tried a linear regresion between Polar Exertion and CPS Tss for 6 months rides and R^2 > 0.9.
PMC with Exertion or Tss looks very similar (actual values are different).

For running I'm using another chart with Exertion data.

I have not figured any method to combine both yet😠