What inspired this wish list request?
when the control limits are user defined, and non symmetric (because of non symmetric distribution) the zone and the WE rules are based on symmetric gaussian distribution
What is the improvement you would like to see?
In case of non symmetric limits (non normal non symmetric distributions) the zones could be defined based on smoothing to adjust the distribution (as in process capability) and using probability and quantiles calculations to "copy" the Shewhart calculations, but it would be complex
Why is this idea important?
Because the zones, and the alarm rate is false