Re: Name of a distribution.



On 30 ec, 13:42, "Diamond, Mark" <d...@xxxxxxx> wrote:
I am looking for the name of a kind of distribution, or alternatively some
indication about how I would search for information on the distribution.
However, because I lack exactly that information, all I can do is describe
how to construct the distribution in which I'm interested.

So here goes ... first a rather poor generalization and then a specific
example.

If you allocate n marbles to n boxes, with the box for each of the n marbles
being chosen uniformly and at random from the n boxes, the what is the
"expected ranked allocation" of marbles in the boxes?

If I allocate three marbles to three boxes, they might be allocated {1,1,1},
with probability 2/9; {0,1,2}, {0,2,1}, {1,0,2}, {1,2,0}, {2,0,1}, {2,1,0},
*each* with probability 1/9; or {0,0,3}, {0,3,0}, {3,0,0}, each with
probability 1/27. Ignoring the permutations and looking at the triplets in
ascending rank order, I get {1,1,1) with probability 1/9, {0,1,2} with
probability 2/3 and {0,0,3} with probability 1/9. In some sense (whether it
is a nice sense or not, I don't know?) the expected rank ordered allocation
is {2/9, 8/9, 17/9} or instead, the normalized proportions are {0.222,
0.889, 1.889}.

Does that distribution have a standard name? If I want to know the similar
proportions for 100 marbles amongst 100 boxes, what distribution should I be
looking for? Answers to those questions or any other comments that might
help elucidate what I'm looking or draw my attention to parallel problems at
would be much appreciated.

Cheers,

Mark Diamond

Your problem is the base of combinatorics. The generating function (a
+ b + c)^3 gives 27 strigs. Between them 3 (aab, aba, baa). These are
reduced to a^2b, if you recognize only the number of marbles in boxes.
The partitions classify all strings as x^2y.
kunzmilan

.



Relevant Pages

  • Re: Name of a distribution.
    ... indication about how I would search for information on the distribution. ... If you allocate n marbles to n boxes, with the box for each of the n marbles ...
    (sci.stat.math)
  • Re: Name of a distribution.
    ... indication about how I would search for information on the distribution. ... If you allocate n marbles to n boxes, with the box for each of the n marbles ...
    (sci.stat.math)
  • Name of a distribution.
    ... I am looking for the name of a kind of distribution, ... If you allocate n marbles to n boxes, with the box for each of the n marbles ...
    (sci.stat.math)
  • Re: Sword of Caine Common Distribution
    ... i added the q's for my 2 boxes as well. ... chris, i hope this helps. ... ink, and a 3-way tie for nocturn, notorious brutality, and psychic ... i'm satisfied that the distribution of commons is pretty friggin ...
    (rec.games.trading-cards.jyhad)
  • Re: 3rd Ed repartition ?
    ... of the European Mathematical Society. ... boxes needed to complete a set in a best-case scenario). ... Fabio said that under ANY distribution (implicitly: ... Fabio said, as quoted in the post linked above, "And IIRC, 400-card ...
    (rec.games.trading-cards.jyhad)