home *** CD-ROM | disk | FTP | other *** search
- Path: sparky!uunet!spool.mu.edu!wupost!waikato.ac.nz!comp.vuw.ac.nz!canterbury.ac.nz!math!wft
- Newsgroups: sci.math
- Subject: Some BIG and SMALL sets of integers.
- Message-ID: <1992Sep11.171806.778@csc.canterbury.ac.nz>
- From: wft@math.canterbury.ac.nz (Bill Taylor)
- Date: 11 Sep 92 17:18:04 +1200
- Distribution: world
- Organization: Department of Mathematics, University of Canterbury
- Nntp-Posting-Host: math.canterbury.ac.nz
- Lines: 73
-
- The thread on the "size" of various sets of positive integers, reminded me of an
- amazing result, which I read in "A Number for Your Thoughts", by S.P.Richards.
- ---
-
- Call a subset of N, "BIG", if the sum of reciprocals of its elements is
- infinite, and "SMALL", if finite. So any set with positive "density", as
- defined in recent posts, is BIG; as can be proved without too much trouble.
-
- But not vice versa. The set of primes is zero-density, but is still BIG;
- whereas the set of powers-of-2 is zero-density and SMALL.
-
- [ Extraneous query: are sets with no limiting density necessarily "big" ? ]
- ---
-
- Anyway, Richards now looks at the decimal representation (groan) of integers,
- and asks, how many have the "correct" number of "1"s. i.e. how many
- integers have *exactly* one tenth of their digits being a "1".
- As it were, "1-average" numbers. 6733310002 is an example.
-
- At first sight it doesn't seem like many, as for a start the integer would
- have to have a multiple-of-10 number of digits. But there are quite a few
- of these, in fact, and with a bit of work one can prove that the set of
- integers with exactly 1/10 of their digits being "1" is BIG.
-
- Of course there is nothing special about "1"; any other named digit would
- easily give the same result. Now suppose we look at the proportions of two
- named digits, say "1" and "2" (again, exactly which two is irrelevant); and
- ask similarly, how many integers have exactly 1/10 of their digits being "1"
- *and* 1/10 of their digits being "2". What we might call "1-2-average" numbers.
-
- Again, this time with quite a bit more work, which I will leave as an excercise
- for the reader, :) one can show that the set of "1-2-average" numbers is
- a BIG set.
-
- So we find that there are "a lot" of 1-digit-average and 2-digit-average numbers.
-
- Obviously, we ask the same about "3-digit-average" numbers, and so on all
- the way up to 9 (equivalently, 10); and it seems obvious that they should all
- be BIG sets of numbers as well.
-
- BUT NO !!!!!!!
- ~~~~~~
- Amazingly, as soon as we go to "3-digit-average" numbers, these are now
- only a SMALL set of numbers. (Again the proof left to the reader.)
- That is, the set of numbers that have exactly 1/10 of their digits being
- a "1" *and* 1/10 being a "2" *and* 1/10 being a "3", is a SMALL set !
- And as every 4-digit-average set is a subset of some 3-digit-average set,
- these (and higher ones) are all SMALL sets as well.
-
- IMHO, this is a truly startling result. One could believe that there might
- be a difference between 1-digit-average and the rest, maybe; or between
- 9-digit-average and those below, perhaps. But that the distinction should come
- between *2* and *3* ; that is really weird !
- ~~~~~~~~~~~~
- ---
- As a final incidental remark: changing the base from 10 to any other base
- from 2 upwards, doesn't essentially change things. This is especially intriguing
- for base 2 and base 3, as then we get that the set of numbers with the "correct"
- proportion of *all* digits is a BIG set !
-
- -------
- A coda. Perhaps those familiar with the recurrence/transcience of
- random walks on lattices in 2/3 dimensions might not find this *quite* so
- startling; and if you search for the proofs of the above results, you may
- find that they hold, "for the same reason" as do the random walk results.
-
- -------------------------------------------------------------------------------
- Bill Taylor wft@math.canterbury.ac.nz
- -------------------------------------------------------------------------------
- Free will- the result of chaotic amplification of quantum events in the brain.
- Galaxies- the result of chaotic amplification of quantum events in the big bang.
- --------------------------------------------------------------------------------
-
-