W
I have been thinking of ...
Calculate Cpk based on the most recent 25 data / measurements and then record that Cpk value. Then re-calculate Cpk every time another measurement happens.
Then plot those Cpk's (Y-axis) as a function of the measurement event (increment the X-axis by 1 unit).
Has anyone found this to be a useful or insightful technique or graph?
Or possibly, is this technique a statistical confusion without merit?
Thank you for your time and thoughts.
Calculate Cpk based on the most recent 25 data / measurements and then record that Cpk value. Then re-calculate Cpk every time another measurement happens.
Then plot those Cpk's (Y-axis) as a function of the measurement event (increment the X-axis by 1 unit).
Has anyone found this to be a useful or insightful technique or graph?
Or possibly, is this technique a statistical confusion without merit?
Thank you for your time and thoughts.
Of course, that notion is not always true anyway...but this really nails it!