Re: YOU ARE WRONG Jack Tomsky

YOU ARE WRONG Jack Tomsky


Since 1930´s we know (those that are not plain CLOWNS
as you are, Jack Tomsky) that the empirical
distributions CONVERGE to the PDF uniformly.


This statement has not been known since the 1930s. In fact, it is still not known today except by Luis A. Afonso, for the simple reason that it is not true. The empirical distributions do not converge to the PDF uniformly. What is true is that the empirical distribution converges uniformly a.s. to the population cdf.

Jack


YOUR repeated ad nausea that CRITICAL VALUES (and
acceptance intervals) were not possible to attain
clarifies well what is your expertise in Statistics.
YOU have to up-date SEVENTY YEARS your statistical
background: a Homeric HOMEWORK


THE DISPLY OF DISGRACEFUL IGNORANCE AND PREJUDICE


1) Date: Mar 29, 2007 7:32 PM
Author: Jack Tomsky
Subject: Re: WHY IS SOUND to get C.V.´s from Mont
Carlo

These "confidence intervals" are not equal to those

***************************

Date: Mar 28, 2007 9:28 AM
Author: Jack Tomsky
Subject: Re: Afonso's little tantrum...

The Afonso Exact Confidence Intervals have an unusual
property. Take the "99% confidence interval" for
N=15, for instance. If the true parameter value is
between -0.655 and 0.655, then the probability of
coverage of parameter is 100%. If the true parameter
is either less than -0.655 or greater than 0.655,
then the probality of coverage is 0%.Thus, the actual
confidence level, which is the inf of the coverage
probabilities is 0%, not 99% as stated. Jack

*******************************
Date: Mar 28, 2007 11:12 AM
Author: Jack Tomsky
Subject: Re: Afonso's little tantrum...

The Afonso Exact Confidence Intervals have terrible
properties with the actual confidence level being 0%.
I just wanted to make sure that no one would ever
r use the Afonso Exact Confidence Intervals. Jack

*******************************

Date: Mar 29, 2007 10:42 AM
Author: Jack Tomsky
Subject: Re: Afonso's little tantrum...

Take his "99% confidence interval" with N=15, for
instance. This was reported as [-0.665, 0.665] for
all samples. If the true parameter value is 0.700,
then the probability of the parameter lying with the
interval is zero. Jack
*****************************


THE COMPLETE IGNORANCE from Jack Tomsky (Stanford Ph.
D. 1974) EVEN OF THE FIRST PTINCIPLES OF STSTISTICAS
THEORY IS SUFFICIENTLY DOCUMENTED.


What you've documented is that I was able to expose the poor coverage performance by the Afonso Exact Confidence Intervals. No one should have any confidence in it.



(What was the Jury that was so conned by the *guy*, I
dare. Jack should not be approved at first exam of
the first course he attended, of course).



If the jury were alive today, they would be appalled that someone could propose such a terrible performing method as the Afonso Exact Confidence Intervals.

Jack


__________licas (Luis A.
.

Relevant Pages