Re: Lilliefors Test : 40 years

***************
This is the 36th anniversary of a paper I wrote with
the late Frank McNolty. J.L. Tomsky and F. McNolty
?Some Properties of Special-Function Distributions",
Sankhya, Vol.34, Part 3, 1972, pp. 251-264. The paper
gives a unified treatment of distributions by
representing special-function distributions as
mixtures of standard distributions. It includes
certain noncentral distributions as special cases.
I'm surprised that Afonso is not celebrating this
anniversary. Jack (moderator)
******************************

My response:
Some OLD AGED people loose the sense of healthy
lucidity about them has got. And when they state
pompously they made important contributions to
science in his miserably professional life the case
is uppermost serious and allowable to be clinically
taken into account.. We ask: Who are Tomsky and
McNolty in Statistics World? They are NOBODY, surely.
(IN THIS WORLD ALL IS RELATIVE).
In what concerns Jack Tomsky it’s not an isolate
episode: unashamedly he associate his name to a
mathematical result he never had notice yet and that
he clearly is unable to get the super-important
consequences for all Computational Statistics: the
Dvoretzky - Kiefer- Wolfowitz theorem. DOCTORS BE
AWARE: YOU ARE DEALING WITH A RECURRENT MANIA!
Luis Amaral Afonso (The Moderator Destroyer)
******************************************
Lilliefors Test : 40 years

It is for me singularly pleasant to remember the 40
years paper on a first attempt to test statistical
hypotheses based on the Empirical Distribution
Function (via Monte Carlo simulations).
Manly because its usefulness and suitability,
recognized all over the world, this exceptional event
cannot be passed unnoticed.
************************
LILLIEFORS, H.W. On the Kolmogorov-Smirnov test for
normality with mean and variance unknown. J. Am.
Stat. Assoc., v.62, p.399-402, 1967.
******************************

Luis Amaral Afonso



Even praising his favorite paper, Afonso's message is filled with misinformation. Kolmorov in 1933 developed a test for testing the simple hypothesis for a completely specified cdf based on the empirical cdf. Thirty-four years later, Lilliefors simply extended it a composite hypothesis for the normal family.

Lilliefors' unsophisticated paper is too complicated for Afonso to understand. In the past, he falsely described the Lilliefors tables as being confidence intervals rather than quantiles of the test statistic.

He also finds simple arithmetic to be too complicated. 2008 - 1967 = 41, not 40. No wonder he couldn't understand my paper.

Jack (moderator)
.