“…From distribution theory, the 2nd and 3rd central moments of a distribution function, which represent the variance and skewness respectively, are given by: (9) From Fig.11 it is apparent that three central moments may be required to estimate the delay with less relative error. Using the Elmore delay as a dominant time constant scales the delay approximation by 0.7 and can help the relative errors somewhat, but not sufficiently in all cases, and with some of the estimates becoming optimistic [10].…”