Normalized Residuals Chart

[Marco]

Hello,

Iam using GC 3.6 with R. In the graph User ODPM i see a Residuals chart.

On the x as i see predicted power but not the time. Iam looking for the same chart as WKO5, where the x give the time. From 1 sec to the right. So the other way than its now.

Is that possible?

[Mark Liversedge]

I don't know WKO5, but the residuals are the difference between predicted and actual *y* values- Power.

Maybe show us the WKO5 chart and lets see what its doing ?

Mark

[Remco Boerwinkel]

Hi Marco,

You may take a look at the normalized residual for testing chart I created a while ago (this one is in python) IIRC I uploaded it to the clouddb.

It finds a short medium and long duration power you should target.

Hope's this fits your need (you might convert it to an new R Chart)

Kind regards,

Remco

Op ma 1 jun. 2020 10:42 schreef Marco <jhoge...@gmail.com>:

--
_______________________________________________
Golden-Cheetah-Users mailing list
golden-che...@googlegroups.com
http://groups.google.com/group/golden-cheetah-users?hl=en
---
You received this message because you are subscribed to the Google Groups "golden-cheetah-users" group.
To unsubscribe from this group and stop receiving emails from it, send an email to golden-cheetah-u...@googlegroups.com.
To view this discussion on the web visit https://groups.google.com/d/msgid/golden-cheetah-users/bd022266-cf3b-4c99-b097-e0df8795ebcd%40googlegroups.com.

[Mark Liversedge]

Ah ok, I found an example on the web, a residuals plot of variance by duration; will update and send here later today

Mark

[Marco]

Hello Mark, sorry for my late reply. Here is the residuals chart from WKO5.

In WKO5 chart only gives time and the difference between predicted and actuals in % i tink...not a given power.

The chart is also in the opposited way. 

Op maandag 1 juni 2020 16:01:21 UTC+2 schreef Mark Liversedge:

[Marco]

Hello Remco,

I dont have Python. I have download it, but it still doesnt work at the moment..

Op maandag 1 juni 2020 13:43:09 UTC+2 schreef RB:

Hi Marco,

You may take a look at the normalized residual for testing chart I created a while ago (this one is in python) IIRC I uploaded it to the clouddb.

It finds a short medium and long duration power you should target.

Hope's this fits your need (you might convert it to an new R Chart)

Kind regards,

Remco

Op ma 1 jun. 2020 10:42 schreef Marco <jhoge...@gmail.com>:

Hello,

Iam using GC 3.6 with R. In the graph User ODPM i see a Residuals chart.

On the x as i see predicted power but not the time. Iam looking for the same chart as WKO5, where the x give the time. From 1 sec to the right. So the other way than its now.

Is that possible?

--
_______________________________________________
Golden-Cheetah-Users mailing list
golden-che...@googlegroups.com
http://groups.google.com/group/golden-cheetah-users?hl=en
---
You received this message because you are subscribed to the Google Groups "golden-cheetah-users" group.

To unsubscribe from this group and stop receiving emails from it, send an email to golden-cheetah-users+unsub...@googlegroups.com.

[Mark Liversedge]

On Monday, 1 June 2020 16:16:56 UTC+1, Marco wrote:

Hello Mark, sorry for my late reply. Here is the residuals chart from WKO5.

No worries.

In WKO5 chart only gives time and the difference between predicted and actuals in % i tink...not a given power.

The chart is also in the opposited way. 

Yeah, there is nothing wrong with that at all, just another way of visualising them. Attached a chart that does the same (its a user chart not R btw).

Attached a fit that uses the ODPM against the first 1200s of data, to mimic WKO5's new "improved" model fitting method.

Be curious to see how they compare with the same data :)

Mark

[Marco]

Thnx Mark, look like it's work. I have make the size max 3600 seconds, but still at the beginning to much squashed.

Nice side effect is the predicted CP is the same as predicted FTP in WKO5. Luck?

Op maandag 1 juni 2020 17:40:42 UTC+2 schreef Mark Liversedge:

[Marco]

Sorry here the attach

Op maandag 1 juni 2020 17:40:42 UTC+2 schreef Mark Liversedge:

On Monday, 1 June 2020 16:16:56 UTC+1, Marco wrote:

[Mark Liversedge]

On Monday, 1 June 2020 17:20:53 UTC+1, Marco wrote:

Thnx Mark, look like it's work. I have make the size max 3600 seconds, but still at the beginning to much squashed.

Nice side effect is the predicted CP is the same as predicted FTP in WKO5. Luck?

Probably. Fitting to unfiltered MMP isn't great, especially out to an hour, where I'm guessing you have no maximal effort?

But glad the chart is useful for you.

Cheers

Mark