Path: utzoo!utgpu!jarvis.csri.toronto.edu!mailrus!uflorida!ukma!tut.cis.ohio-state.edu!osu-cis!att!ihlpl!kat3
From: kat3@ihlpl.ATT.COM (Craig)
Newsgroups: comp.lang.c
Subject: standard deviation
Message-ID: <9849@ihlpl.ATT.COM>
Date: 27 Mar 89 01:35:48 GMT
Reply-To: kat3@ihlpl.UUCP (Craig,R.J.)
Distribution: usa
Organization: AT&T Bell Laboratories - Naperville, Illinois
Lines: 15

> 	Sure; and you could just as easily have picked another data
> set that produced a correct result from the onepass program, and an
> incorrect result from twopass.  Which is why, as has been pointed out
> before, neither algorithm should be used in a general standard
> deviation program.

I don't think you can find such a data set.  The given onepass algorithm
is just "BAD."  There are other algorithms which are "better" in the
sense that they are faster, but all "good" algorithms for the computation
use "centering" to increase accuracy.

If you can provide such a data set, I would be very interested.


					Bob Craig