Relay-Version: version B 2.10 5/3/83; site utzoo.UUCP
Posting-Version: version B 2.10.1 6/24/83; site charm.UUCP
Path: utzoo!watmath!clyde!burl!ulysses!mhuxl!mhuxj!mhuxi!charm!mam
From: mam@charm.UUCP (Matthew Marcus)
Newsgroups: net.misc
Subject: Re: Horses vs. Currant Buns: a query
Message-ID: <279@charm.UUCP>
Date: Mon, 19-Mar-84 14:23:51 EST
Article-I.D.: charm.279
Posted: Mon Mar 19 14:23:51 1984
Date-Received: Tue, 20-Mar-84 05:55:35 EST
References: <712@seismo.UUCP>
Organization: Physics Research - AT&T Bell Labs MH
Lines: 43

Seismo!flinn divides people into two classification, which I will call H and C,
for reasons clear in the basenote.  He then asks why most married couples comprise
an H and a C, rather than two H's or two C's.  Since he posted this to net.misc,
I assume he thinks there is some non-statistical reason for this.  I will show
that his observation can be explained by random association given certain
simple assumptions:

	Let x be the probability that a man is in class C.  Let y be the probability
of a woman being a C.  Now we make a table:

Couple		probability
Both C		xy
Both H		(1-x)(1-y)
M=C,W=H		x(1-y)
M=H,W=C		y(1-x)

Then, let D=Pr(M<>W)-Pr(M=W), where M,W refer to male, female members of
a couple, relational operators refer to membership in the C or H classes,
and Pr(c) is the probability of condition c being true. Seismo!flinn's
observation is that D>0.  Now, with the above table, we find that
D=2x+2y-4xy-1
Substitute x=(w+1)/2, y=(z+1)/2, and get
D = -wz  .
which shows that if x-1/2 and y-1/2 are of different signs, then flinn's
observation follows.  There is an assumption here that there is no correlation
between one's own type and the type you will get married to. The final result is:

If {seismo!flinn's observation}
  THEN	{
	If {random chance applies to marriages viz. H, C classes}
	   THEN	{
		one sex shows an excess of C and the other an excess of H
		}
	   ELSE {
		C-type men prefer H-type women, or vice-versa
		}
	}
  ELSE	{
	Much time has been wasted
	}


		{BTL}!charm!mam  "why isn't this in net.math?"