ABSTRACT
Statistical design and control both require proof from businesses improving suddenly then sustaining. More information has been provided by now about why there is no mathematics to predict whether a particular completed statistical design will, or will not, improve a process, until it does. Statistical control recognizes there is also no mathematics to explain (or predict) why a process improves (or by how much) when it is stabilized. e method is therefore designed to exploit the real world, in real-time, as it unfolds. ere is no theory to calculate sample sizes except roughly, in statistical design, because the variation oen tightens. Comparisons of variation before and during testing are given in several cases.
