Re: OT - simple (not for me) probability question



David (Ross) confirms:
Jack got this one correct (twice!).

He wasn't alone ... Eyes glaze over when the indistinguishable urns
and balls appear, but I may be able to illuminate the problem a
little.

It's usually easier to compute simpler, related results first, and
then build upon them. In this case, one of the simplest results is the
all-red probability, 1/1024 (by inspection). Indistinguishability
plays a role with the next result. The first ball is not red 3/4 of
the time, and the rest are red 1/256 of the time, but this case is
indistinguishable from the one where the only non-red ball comes from
the second urn. In other words, there are five ways to achieve this
result so the probability is 5 x 3/4 x 1/256 = 15/1024.

The more complicated cases are best addressed by studying the effect
of indistinguishability, but those of us who've given up on acquiring
that intuition may find it easier to look at the corresponding four-
urn problem instead. Here, the simple results indirectly produce a
complete solution: none-red = 81/256, one-red = 108/256, all-red =
1/256, three-red = 12/256, and two-red must be the fraction needed to
get these to sum to 1, i.e. 54/256.

Restoring the fifth urn ... if the first ball is red, we need two
more, so the probability is 1/4 x 54/256 = 54/1024. If the first ball
is not red we need three more, so the probability is 3/4 x 12/256 =
36/1024. Note that indistinguishability is already factored into these
results. Thus, the total three-red probability is 90/1024 (as before).


Felix

.



Relevant Pages

  • Re: Religion VS Science
    ... the probability of the hypothesis is 50%. ... >> ball randomly from the urn. ... but I do understand that the probability changes). ...
    (misc.fitness.weights)
  • Re: Placing Balls in Urns and Expected values
    ... probability and we supposed the balls are ... Probability for 1 urn to be empty is ^1 ... Probability for n urns to be empty is ^n ... A ball is placed randomly in one of the n urns with an equal probability. ...
    (sci.math)
  • Re: Balls in Urns
    ... is placed in any particular urn with probability 1/n. ... If we label the balls, the total number of outcomes is n^m; ... the outcomes with no more than one ball in any urn is n...(n-m ...
    (sci.math)
  • Re: Balls in Urns
    ... is placed in any particular urn with probability 1/n. ... Note that these outcomes are equally likely. ... the outcomes with no more than one ball in any urn is n...(n-m ...
    (sci.math)
  • Re: Uncountably of Super Infinite Sets
    ... Inside each urn is a ball with the same ... It is -- especially if one then tries to add a ball to an urn by ... of bijections between sets and their proper subsets. ... Outside of a dog, a book is man's best friend. ...
    (sci.math)