Gross Power Released/Gross Power Absorbed

[Valvero]

Is there in the cloud or do you have a chart to measure the GPR / GPA of pedaling?

[Valvero]

None?  I have been looking for the definitions of each metric. For example: GPR (right leg) is defined as the product of power times the effectiveness of torque [GPR=P*T_eff]. GPA (right leg) is defined as the product of power minus GPR [GPA= P-GPR]. Similarly for the left leg.

But I don't know how to configure it.

Can anyone help?

[Ale Martinez (Please don't email or cc me)]

To get you started GPR left could be something like this: samples(POWER)*samples(BALANCE)/samples(LEFTEFFECTIVENESS)

[Valvero]

Thanks Ale,
I've been trying to do it, but it doesn't seem to work

[Ale Martinez (Please don't email or cc me)]

It looks like your chart has only one Axis, check the Axes tab, this is the main issue.

Additionally init and sample methods are not doing anything useful and can be removed.

For example, to chart GPRL over time with annotation for average value:

To compute GPAL in other curve you can reuse GPRL curve values without recomputation: yy <- curve(GPRL,y)-samples(POWER)*samples(BALANCE)/100;

[Valvero]

Now it seems to work fine, but ... if I want to represent watts ... versus total watts ... how can I do it?

A graphic of this style

or against time, but in the same style as the PDC

[Ale Martinez (Please don't email or cc me)]

El lunes, 19 de octubre de 2020, 13:02:50 (UTC-3), Valvero escribió:

Now it seems to work fine, but ... if I want to represent watts ... versus total watts ... how can I do it?

As you attempted before, but give different names to the axes, like Watts and Total Watts

 

A graphic of this style

or against time, but in the same style as the PDC

 https://github.com/GoldenCheetah/GoldenCheetah/wiki/UG_Special-Topics_Formula-Syntax-and-Expressions#generating-a-mean-maximal-curve-from-other-data

[Valvero]

Now it has returned to seem not to work

[Ale Martinez (Please don't email or cc me)]

El lunes, 19 de octubre de 2020, 13:16:23 (UTC-3), Valvero escribió:

Now it has returned to seem not to work

I don't know what you are talking about.

 

What I see is what looks like a scatter chart with lines, which can be easily avoided changing the line style to none.

 

[Valvero]

Currently I have it as you indicated to me 

But one of the expressions that I don't quite understand is the power performed by each leg. I guess it comes from samples(POWER)*samples(BALANCE). But how do we differentiate the right leg from the left?

Pd: I don't know why but it won't let me attach the graphic file.

[Ale Martinez (Please don't email or cc me)]

El martes, 20 de octubre de 2020, 4:36:38 (UTC-3), Valvero escribió:

Currently I have it as you indicated to me 

But one of the expressions that I don't quite understand is the power performed by each leg. I guess it comes from samples(POWER)*samples(BALANCE). But how do we differentiate the right leg from the left?

Think about what BALANCE series represents  

[Valvero]

Better?

[Ale Martinez (Please don't email or cc me)]

El martes, 20 de octubre de 2020, 5:23:25 (UTC-3), Valvero escribió:

Better?

Yep

[Valvero]

Great,

I would like to share this graphic in the cloud, but I would like to know if it is all right and makes sense.

One thing I don't quite understand is the purple dots that appear diagonally.  What is it?

[Ale Martinez (Please don't email or cc me)]

El martes, 20 de octubre de 2020, 13:56:03 (UTC-3), Valvero escribió:

Great,

I would like to share this graphic in the cloud, but I would like to know if it is all right and makes sense.

Feel free to upload, it can be interesting for other users, just add a description and an indication it requires v3.6.

Moving the labels to the right may avoid the need for horizontal scrolling.

 

One thing I don't quite understand is the purple dots that appear diagonally.  What is it?

GPAL points on diagonal line likely correspond to zero GPRL near the x-axis, this may help to debug.

If you attach the gchart and fit files, I could take a look.

[Valvero]

I don't know why, but it won't let me attach the .gchart file

[Ale Martinez (Please don't email or cc me)]

El martes, 20 de octubre de 2020, 15:09:43 (UTC-3), Valvero escribió:

I don't know why, but it won't let me attach the .gchart file

It's weird, I can attach .gchart files using the web interface.

Anyway, probably some points need to be filtered out to avoid this effect like extremes with BALANCE=0, 100 or something like this.

See https://github.com/GoldenCheetah/GoldenCheetah/wiki/UG_Special-Topics_Formula-Syntax-and-Expressions#selecting-and-filtering-a-vector

[Valvero]

xx <- samples(POWER);

yy <- curve(GPRL,y)-samples(POWER)*samples(BALANCE)/100;

yy <-yy[>0];

annotate(label, "GPAL avg", round(mean(yy)));

??

[Ale Martinez (Please don't email or cc me)]

El martes, 20 de octubre de 2020, 16:58:53 (UTC-3), Valvero escribió:

xx <- samples(POWER);

yy <- curve(GPRL,y)-samples(POWER)*samples(BALANCE)/100;

yy <-yy[>0];

annotate(label, "GPAL avg", round(mean(yy)));

??

That's not even syntactically correct.

Anyway since we don't know which the problem is, it would be better to add a SECS series to find out to which data points they correspond in Raw Data and try to understand the cause looking there.

[mlt.m...@googlemail.com]

Something I did wrong with the first post.

Here again my assumption in addition.

For the left leg. Very roughly.

Drive power left=Power*BALANCE*LEFTEFFECTIVENESS
Is the power that moves you forward.
BALANCE indicates the performance distribution and LEFTEFFECTIVENESS how efficient the process is. The efficiency (LEFTEFFECTIVENESS) is not always 100%, the result of Power*BALANCE*LEFTEFFECTIVENESS must be < Power*BALANCE.

Power loss left=Power*BALANCE*(100-LEFTEFFECTIVENESS)
Is the power that works against the propulsion on. That's why I have shown it negatively in the chart.

Both formulas together give the power for the left leg.
Power*BALANCE*LEFTEFFECTIVENESS+Power*BALANCE*(100-LEFTEFFECTIVENESS)

The sample whether the assumption is correct.
Power*BALANCE=Power*BALANCE*LEFTEFFECTIVENESS+Power*BALANCE*(100-LEFTEFFECTIVENESS)

Then I filtered all values at BALANCE which are negative. There I do not know, if this makes sense.

[Ale Martinez (Please don't email or cc me)]

El miércoles, 21 de octubre de 2020, 11:10:11 (UTC-3), mlt.m...@googlemail.com escribió:

Something I did wrong with the first post.

Here again my assumption in addition.

For the left leg. Very roughly.

Drive power left=Power*BALANCE*LEFTEFFECTIVENESS
Is the power that moves you forward.
BALANCE indicates the performance distribution and LEFTEFFECTIVENESS how efficient the process is. The efficiency (LEFTEFFECTIVENESS) is not always 100%, the result of Power*BALANCE*LEFTEFFECTIVENESS must be < Power*BALANCE.

Power loss left=Power*BALANCE*(100-LEFTEFFECTIVENESS)
Is the power that works against the propulsion on. That's why I have shown it negatively in the chart.

Both formulas together give the power for the left leg.
Power*BALANCE*LEFTEFFECTIVENESS+Power*BALANCE*(100-LEFTEFFECTIVENESS)

The sample whether the assumption is correct.
Power*BALANCE=Power*BALANCE*LEFTEFFECTIVENESS+Power*BALANCE*(100-LEFTEFFECTIVENESS)

I thought you were talking about GPR/GPA as defined by Andrew Coggan so I suggested formulas based on: https://irp-cdn.multiscreensite.com/dc479d29/files/uploaded/WKO4%20Pedaling%20Metrics.pdf, your new formulas are something else.

 

Then I filtered all values at BALANCE which are negative. There I do not know, if this makes sense.

That's is the kind of bogus data likely producing the weird line in your previous chart and it should be filtered.

[mlt.m...@googlemail.com]

Do not understand the formula from the PDF.

If I understand the PDF correctly:
GPR=P+
GPA=P-
--> GPR+GPA=total Power
--> (GPR+GPA)/total Power=<1

When I calculate this with numbers I come to this:
total power=268 W; balance=0.54;
torque effectiveness L= 95; torque effectiveness R = 78

GPRL=(total Power *balance*100)/torque effectiveness L
GPRL=(268*0.54*100)95=152.34 W

GPRR=(total power *(1-balance)*100)/torque effectiveness R
GPRR=(268*(1-0,54)*100)78=158,05 W

GPAL=(total power *balance*100)/torque effectiveness L-total power *balance
GPAL=(268*0,54*100)95-268*0,54=7,61 W

GPAR=(total power *(1-balance)*100)/torque effectiveness R-total power *(1-balance)
GPAR=(268*(1-0,54)*100)78-268*(1-0,54)=34,77 W

GPR+GPA=GPRL+GPRR+GPAL+GPAR=152,34+158,05+7,61+34,77=352,77 W

(GPR+GPA)/total Power=352,77/268=1,32
--> 132 %

If I see GPR as usable power and GPA as non-usable power and total power as supplied power.
I cannot make the GPR+GPA greater than total power.

[Ale Martinez (Please don't email or cc me)]

El miércoles, 21 de octubre de 2020, 16:17:20 (UTC-3), mlt.m...@googlemail.com escribió:

Do not understand the formula from the PDF.

If I understand the PDF correctly:
GPR=P+
GPA=P-
--> GPR+GPA=total Power
--> (GPR+GPA)/total Power=<1

When I calculate this with numbers I come to this:
total power=268 W; balance=0.54;
torque effectiveness L= 95; torque effectiveness R = 78

GPRL=(total Power *balance*100)/torque effectiveness L
GPRL=(268*0.54*100)95=152.34 W

GPRR=(total power *(1-balance)*100)/torque effectiveness R
GPRR=(268*(1-0,54)*100)78=158,05 W

GPAL=(total power *balance*100)/torque effectiveness L-total power *balance
GPAL=(268*0,54*100)95-268*0,54=7,61 W

GPAR=(total power *(1-balance)*100)/torque effectiveness R-total power *(1-balance)
GPAR=(268*(1-0,54)*100)78-268*(1-0,54)=34,77 W

GPR+GPA=GPRL+GPRR+GPAL+GPAR=152,34+158,05+7,61+34,77=352,77 W

(GPR+GPA)/total Power=352,77/268=1,32
--> 132 %

If I see GPR as usable power and GPA as non-usable power and total power as supplied power.
I cannot make the GPR+GPA greater than total power.

I think it is a sign confusion, look at the not below Figure 11-1:

It looks like GPA is defined to be positive, the absolute value of P-, so GPR-GPA=Total Power

[Ale Martinez (Please don't email or cc me)]

Since the chart you upload to CloudDB is in line with the PDF definitions I edited the description and moved it to curated section, it is a nice example of the new User Charts in v3.6 and other users may find it useful as is or to adapt to their taste, thanks for the contribution.

[Valvero]

I am looking for information to complement the graph. My idea would be to add the Kurtotic index, which is the ratio between the maximum effective force produced on the pedal and the average effective force in each pedal stroke. The lower the value of this index, the better pedaling technique. But I can't find the mathematical expression.

Do you have some information about this?

[Ale Martinez (Please don't email or cc me)]

El sábado, 24 de octubre de 2020, 16:21:17 (UTC-3), Valvero escribió:

I am looking for information to complement the graph. My idea would be to add the Kurtotic index, which is the ratio between the maximum effective force produced on the pedal and the average effective force in each pedal stroke. The lower the value of this index, the better pedaling technique. But I can't find the mathematical expression.

Do you have some information about this?

No

[Ale Martinez (Please don't email or cc me)]

El sábado, 24 de octubre de 2020, 16:35:07 (UTC-3), Ale Martinez (Please don't email or cc me) escribió:

El sábado, 24 de octubre de 2020, 16:21:17 (UTC-3), Valvero escribió:

I am looking for information to complement the graph. My idea would be to add the Kurtotic index, which is the ratio between the maximum effective force produced on the pedal and the average effective force in each pedal stroke. The lower the value of this index, the better pedaling technique. But I can't find the mathematical expression.

Do you have some information about this?

No

BTW, based on that description it looks like the inverse of Pedal Smoothness: https://support.garmin.com/en-US/?faq=G41ZZf19cK4OhEM7XVkrc8

[Valvero]

I've tried this, but it doesn't seem to show anything

[Ale Martinez (Please don't email or cc me)]

El sábado, 24 de octubre de 2020, 17:55:24 (UTC-3), Valvero escribió:

I've tried this, but it doesn't seem to show anything

At first sight 1/ doesn't seem right since pedal smoothness is expressed as percentage.

Anyway, in more general terms, I think "It doesn't work in my first attempt so I take a picture and post to the forum" is not a sustainable debugging strategy...

[A Evans]

Hi. Just got some powermeter pedals.

Thanks for the curated chart above. Good stuff.

Did anyone ever figure out the formula for Kurtotic Index? Is it just 100/smoothness or is it more than that?

Also has anyone got a mmp chart set up for GPR/GPA, like:

https://groups.google.com/group/golden-cheetah-users/attach/4b0e9c92008e4/Captura2.PNG?part=0.1&view=1

Just thought it would be interesting.

Thanks

[Mark Liversedge]

Kurtosis is just a measure of long tailedness in distributions, sounds fancy tho, so I am guessing its a Coggan thing.

Send me a file with GPR/GPA in it and can create a user chart for that.

Mark