Custom PMC

134 views
Skip to first unread message

Bob McRae

unread,
Apr 30, 2018, 2:41:02 PM4/30/18
to golden-cheetah-users
I have searched the forum & wiki, but not quite found what I am after.  I am interested in creating a custom PMC and daily metrics chart like that described below.  What is the best way to go about this?  Will I need to do my own development, or can I use GC's user defined metrics?
  • Custom PMC:
    • Uses TSS for cycling and Normalized Graded pace for running so that the CTL, TSS, and ATL numbers substantially match TrainingPeaks/WKO
    • Allows changing of the CTL model to be sigmoidal, rather than exponential.  I disagree with the concept that a week of rest results in a ~10% reduction in CTL.  Modeling the CTL with a sigmoidal decline would better represent reality
  • N-day Moving Average Chart
    • I like to track how much TSS, miles, work, etc I do over a 3, 4, or 5-day period as I believe that is applicable to what one can do in a single day.  However, I am not seeing a way to create a metric that works on calendar days; it seems to be based on activities.  While I have activities nearly every day, I'd like to be able to process the metric based on calendar days
Thanks in advance for the replies.

Ale Martinez

unread,
Apr 30, 2018, 3:55:35 PM4/30/18
to golden-cheetah-users
El lunes, 30 de abril de 2018, 15:41:02 (UTC-3), Bob McRae escribió:
I have searched the forum & wiki, but not quite found what I am after.  I am interested in creating a custom PMC and daily metrics chart like that described below.  What is the best way to go about this?  Will I need to do my own development, or can I use GC's user defined metrics?
  • Custom PMC:
    • Uses TSS for cycling and Normalized Graded pace for running so that the CTL, TSS, and ATL numbers substantially match TrainingPeaks/WKO
NGP has not a published operational definition; AFAIK. The closest metric in GC is xPace.
    • Allows changing of the CTL model to be sigmoidal, rather than exponential.  I disagree with the concept that a week of rest results in a ~10% reduction in CTL.  Modeling the CTL with a sigmoidal decline would better represent reality
You could do that in GC using an R (v3.4) or Python (v3.5) chart. 
  • N-day Moving Average Chart
    • I like to track how much TSS, miles, work, etc I do over a 3, 4, or 5-day period as I believe that is applicable to what one can do in a single day.  However, I am not seeing a way to create a metric that works on calendar days; it seems to be based on activities.  While I have activities nearly every day, I'd like to be able to process the metric based on calendar days
if you want to chart that value day-by-day, independently of the presence of activities you would need an R/Python chart, as above.
Custom metrics are computed for activities, not for days, if you want to consider a 3, 4 or 5 day period before an activity you could do that using a Python script inside the custom metric (only v3.5, see the wiki).
Reply all
Reply to author
Forward
0 new messages