Multi activities in multisports fitness trend

[geppoluc]

Bug description: I noticed that the multisport fitness trend only considerds one activity per day even in case more than one activity is available. For example, today I did one cycling activity outdoor (only hr data available) and a shor activity on a trainer indoor (hr and powermeter data available). The TRIMP is the result of only one activity, in my case the shortest one, i.e. the indoor one.

Actual Behavior: Only one activity is considered by Stravistix

Expected Behavior: The multisport fitness trend should consider the two activities in the assesment of the TRIMP, PSS, Fitness and Fatigue, as other software are correctly doing (for instance Golden Cheetah)

[thomaschampagne]

@geppoluc Could provide me a backup of your history to debug?

Thanks

[geppoluc]

Hi Thomas, attached please find the backup of my history. (https://www.dropbox.com/s/20i58p4px0rzlw9/2017.1.29-20.18_v5.4.0.history.json?dl=0)

The two activities are: https://www.strava.com/activities/848672764 (hr and powermeter data) https://www.strava.com/activities/848157602 (only hr data, this activity is not considered in the calculation of fitness, fatigue and form)

I also tried to delete from Strava the first activity (the one with hr and power data) and only then the multisport fitness trend was considering the other activity

Let me know if you need any additional detail Cheers and thanks for your nice addon, Luca

On Sun, 29 Jan 2017 06:47:22 -0800 Thomas Champagne notifications@github.com wrote:

@geppoluc Could provide me a backup of your history to debug?

Thanks

-- You are receiving this because you were mentioned. Reply to this email directly or view it on GitHub: https://github.com/thomaschampagne/stravistix/issues/329#issuecomment-275918293

[thomaschampagne]

@geppoluc Your problem should be fixed :)

Here is a build from CI if you want to test and confirm the fix: https://ci.appveyor.com/project/thomaschampagne/stravistix/branch/develop/artifacts

Thx ;)

Tom

[thomaschampagne]

Mistake from me... should not be fixed finally. I missed something

[thomaschampagne]

@geppoluc fixed for sure yersterday evening.

Get test build here: https://ci.appveyor.com/project/thomaschampagne/stravistix/branch/develop/artifacts Install steps: https://github.com/thomaschampagne/stravistix/wiki/How-to-install-stravistix-build-archive

Tom

[geppoluc]

It looks to be working. I saw you introduced the “Total Stress”.

I am not sure the PSS should be simply summed to the TRIMP, this is something you may know.

What I know is that other software that I am using do not do it. They simply consider either the TRIMP or a stress score assessed relying on the powermeter data. And they do it without mixing the two components which is something extremely interesting for multisport/ multi-bike training.

Thanks,

Luca

Da: Thomas Champagne [mailto:notifications@github.com] Inviato: 31 January 2017 09:43 A: thomaschampagne/stravistix Cc: geppoluc; Mention Oggetto: Re: [thomaschampagne/stravistix] Multi activities in multisports fitness trend (#329)

@geppoluc https://github.com/geppoluc fixed for sure yersterday evening.

Get test build here: https://ci.appveyor.com/project/thomaschampagne/stravistix/branch/develop/artifacts Install steps: https://github.com/thomaschampagne/stravistix/wiki/How-to-install-stravistix-build-archive

Tom

— You are receiving this because you were mentioned. Reply to this email directly, view it on GitHub https://github.com/thomaschampagne/stravistix/issues/329#issuecomment-276305947 , or mute the thread https://github.com/notifications/unsubscribe-auth/AYPR4DIcMfRGpQMcJD3m5PecIxlt11mUks5rXvQHgaJpZM4LwyuB . https://github.com/notifications/beacon/AYPR4N_VJ6Zb12LOC4LfsHi5YEEap38kks5rXvQHgaJpZM4LwyuB.gif

[thomaschampagne]

I'm not mixing PSS and TRIMP for a same activity.

I sum X stress scores for X activities of a day.

X can be PSS (if available) even if you have TRIMP available for the same activity. Else X is TRIMP

[geppoluc]

I see.

I'm not an expert and just new to these topics, but I noticed that for most of my activities (for which hr and power data are concurrently available) the TRIMP score is nearly the double of the PSS score. For instance in this activity (https://www.strava.com/activities/845319807) the TRIMP is 90 and PSS 45.

I do not know if there is anything (which you probably know) in the scientific literature related to the aggregation of the two scores (PSS and TRIMP). In case there is not it would be an interesting topic to be investigated.

As a preliminary (and brutal) attempt, for each athlete, I would define a "substitution rate" based on the scores of old activities:

K=average(TRIMP_old/PSS_old) In my case it would be K~=2

K could be used for the estimation of the Total Stress in case hr/power data is missing.

But maybe I am completely wrong.

Cheers, Luca

On Tue, 31 Jan 2017 05:56:39 -0800 Thomas Champagne notifications@github.com wrote:

I'm not mixing PSS and TRIMP for a same activity.

I sum X stress scores for X activities of a day.

X can be PSS (if available) even if you have TRIMP available for the same activity. Else X is TRIMP

-- You are receiving this because you were mentioned. Reply to this email directly or view it on GitHub: https://github.com/thomaschampagne/stravistix/issues/329#issuecomment-276368961

[thomaschampagne]

Hi luca,

Does your ftp well entered in stravistix? I'm expreriencing same thing on strava fitness graph when use Trimp + PSS.

I also have same diff than you

PSS formula i use: https://github.com/thomaschampagne/stravistix/blob/develop/plugin/options/app/services/FitnessDataService.ts#L111

Which value other software (which one?) are giving you for 29/01 ?

(can't open your private? activity)

Tom

[geppoluc]

Hi Tom,

I checked my settings in Stravistix (FTP = 266 W (estimated from CP20 = 280 W), Max hr = 181, rest hr = 52). To define them I took a 20’ test with a personal trainer, which means that I am quite confident about these values.

I use Golden Cheetah (http://www.goldencheetah.org/), which you probably know and I find to be really complete, maybe too complex for an average (Strava) user. For that same activity (26/1) it gives me TRIMP = 87 (instead of 90 by Stravistix) and TriScore = 48 (instead of PSS = 49 by Stravistix). The two scores are reasonably close to think the difference could be a matter of rounding or normalization. I modified the privacy of the activity (https://www.strava.com/activities/845319807/overview).

What I was noticing is that Golden Cheetah produces two different “Training Stress Balance” charts: (1) for the TRIMP and (2) for the TriScore.

Cheers,

Luca

Da: Thomas Champagne [mailto:notifications@github.com] Inviato: 01 February 2017 09:06 A: thomaschampagne/stravistix Cc: geppoluc; Mention Oggetto: Re: [thomaschampagne/stravistix] Multi activities in multisports fitness trend (#329)

Hi luca,

Does your ftp well entered in stravistix? I'm expreriencing same thing on strava fitness graph when use Trimp + PSS.

I also have same diff than you

PSS formula i use: https://github.com/thomaschampagne/stravistix/blob/develop/plugin/options/app/services/FitnessDataService.ts#L111

Which value other software (which one?) are giving you for 29/01 ?

(can't open your private? activity)

Tom

— You are receiving this because you were mentioned. Reply to this email directly, view it on GitHub https://github.com/thomaschampagne/stravistix/issues/329#issuecomment-276596020 , or mute the thread https://github.com/notifications/unsubscribe-auth/AYPR4G-04pMVOKsgo5Yv__EzdO6ff76hks5rYDzmgaJpZM4LwyuB . https://github.com/notifications/beacon/AYPR4OP7XQxwFGcNUa8jkNUkYl2mY591ks5rYDzmgaJpZM4LwyuB.gif

[thomaschampagne]

@geppoluc Yes i used it in the past.

"(1) for the TRIMP and (2) for the TriScore." Do i am supposed to see some images? if yes can't see them.

In your case on 26/01 stravistix will use a value of 49 for the impulsion if powermeter option activated, else 87. I'm quite confident with calculation of TRIMP or PSS (or TSS trainingpeaks trademarked term)

Quoting this article: http://home.trainingpeaks.com/blog/article/the-science-of-the-performance-manager

Chronic training load, or CTL, provides a measure of how much an athlete has been training (taking into consideration both volume and intensity) historically, or chronically. It is calculated as an exponentially-weighted moving average of daily TSS (or TRIMP, etc.) values.

So it's seems to be PSS or TRIMP, no mix or fusion of them if you have both on same ride.

Maybe this require more documentation? Or someone more experienced in?

[geppoluc]

I know traininpeaks but I have never used it. The problem in this discussion is that I do not know the subject and, for this reason, I’ll read the article carefully later, after work.

Does traininpeaks behave the same as Stravistix? If so (as I suppose), then the problem is me.

I am saying mixing/fusing because CTL is the “exponentially-weighted moving average of daily TSS (or TRIMP, etc.)”. So, if on day 1 you have the TRIMP and on day 2 you have the TSS, you are averaging (summing) the two (different?) objects. Or, as in my case, you have two different activities in the same day, one with TRIMP and the other with TSS, the CTL is the sum of the two.

I’m sorry for this discussion, I do not really want to make you lose time more than I already did.

Da: Thomas Champagne [mailto:notifications@github.com] Inviato: 01 February 2017 15:32 A: thomaschampagne/stravistix Cc: geppoluc; Mention Oggetto: Re: [thomaschampagne/stravistix] Multi activities in multisports fitness trend (#329)

@geppoluc https://github.com/geppoluc Yes i used it in the past.

"(1) for the TRIMP and (2) for the TriScore." Do i am supposed to see some images? i yes can't see them.

In your case on 26/01 stravistix will use a value of 49 for the impulsion if powermeter option activated, else 87. I'm quite confident with calculation of TRIMP or PSS (or TSS trainingpeaks trademarked term)

Quoting this article: http://home.trainingpeaks.com/blog/article/the-science-of-the-performance-manager

Chronic training load, or CTL, provides a measure of how much an athlete has been training (taking into consideration both volume and intensity) historically, or chronically. It is calculated as an exponentially-weighted moving average of daily TSS (or TRIMP, etc.) values.

So it's seems to be PSS or TRIMP, no mix or fusion of them if you have both on same ride.

Maybe this require more documentation? Or someone more experienced in?

— You are receiving this because you were mentioned. Reply to this email directly, view it on GitHub https://github.com/thomaschampagne/stravistix/issues/329#issuecomment-276671610 , or mute the thread https://github.com/notifications/unsubscribe-auth/AYPR4Nsf721JHToCKMFhkROunqkp2Vusks5rYJdugaJpZM4LwyuB .Immagine rimossa dal mittente.