Random Sequencing

[Antti Karttunen]

Olivier Gerard wrote:
> 
> Dear Jon,
> 
> For this kind of musings which borders on the un-mathematical (since the average answer
> of the OEIS does not depend on physics or mathematics but mainly
> on OEIS history and its contributors' humanity, ability and creativity), you
> should do it yourself (number of answers is limited in the public
> version of the OEIS since this is a shared resource, and something
> like 1,2,3 brings something like 8600 potential answers):
> 
>         - download the whole OEIS files (around 46 Mo, 70+ files)
          etc...


Or do a modified Markov chain (upto some degree d) analysis on all the sequences,
and amuse us by showing how "the most expected" OEIS sequence looks like on certain date
of autumn 2002.
For that downloading only http://www.research.att.com/~njas/sequences/stripped.gz
should suffice (5 megs).

With modified I mean that one should probably also record the position
of each term, as at least we would have triplets (pos, term_{in_pos}, term_{in_pos+1})
(for the first degree Markov chain), otherwise I'm afraid we end with
the all-1 sequence, whose A-number I can't find now, because OEIS keeps
returning myriads of Legendre symbol (n,...) sequences for "1,1,1,1,1,1,1,1,...".

Indeed, sorting the results by ascending A-number as suggested by many
would be a very good idea.

Frivolously & uncoherently yours,

Antti Karttunen