Subject: Re: 3n+1 problem
From: rpw3@rpw3.org (Rob Warnock)
Date: Wed, 22 Apr 2009 20:23:50 -0500
Newsgroups: comp.lang.lisp
Message-ID: <JvCdnfWjA-e7XnLUnZ2dnUVZ_rqdnZ2d@speakeasy.net>
Madhu  <enometh@meer.net> wrote:
+---------------
| Willem Rein Oudshoorn <m2mya8ojaw.fsf@Silver.oudshoorn.nl> :
| | A(m) ~  7.228262518959627 * log2(m)
...
| | bits in m.
...
| 
| If I understand it right, the crucial assumptions on which this hinges
| are 1) that repetition of numbers in the sequence and the possible
| existence of cycles for large n will not impact the result of the
| analysis, and 2) odd and even numbers at any step of the sequence are
| equally likely.  (Both assumptions seem OK to me)]
+---------------

In the original formulation of the problem, Assumption #2 was *definitely*
incorrect, since (a) odd numbers can *only* be followed by even numbers,
and (b) even numbers can be followed by up to log2(n) even numbers.
What I haven't worked through yet is whether this strong bias towards
even numbers in the original problem follows through into Oudshoorn's
reformulation...


-Rob

-----
Rob Warnock			<rpw3@rpw3.org>
627 26th Avenue			<URL:http://rpw3.org/>
San Mateo, CA 94403		(650)572-2607