Re: Please help me to find a mistake here


valery wrote:
matt271829-news@xxxxxxxxxxx wrote:
Some questions...

Must A', A'', B and B'' all be distinct? For example, if you were
working with 2^3 rather than 2^32 then would 000, 001, 011 count as two
pairs or not?

actually 000, 001, 011 could count as two pairs (001, 000) and (001,
011), in case if it makes it easier to count probabilities. However
this is degenerate case that is not very interesting for the purposes
of my experiment.


I was trying to find an exact formula - in which case all these minor
points matter - but it doesn't seem to be particularly easy.
Incidentally, continuing in picky mode, in your original post you said
the numbers were 0,1,...,2^32. Did you actually mean 0,1,...,2^32-1?

Anyway, FWIW, making some simplifying assumptions (that seem entirely
reasonable since 71 is so much smaller than 2^32), I make the
probability of event E equal to about 1.76E-10. I think this is
slightly different from James Waldby's answer. I did it in a slightly
more elaborate way, but with these particular numbers I would have
thought the answers should differ by only an imperceptible amount. I'm
not sure at the moment why the answers differ as much as they do, but
if I get a following wind I might try to pick through it and figure it
out.

.

Relevant Pages

  • Re: Please help me to find a mistake here
    ... in case if it makes it easier to count probabilities. ... Incidentally, continuing in picky mode, in your original post you said ... thought the answers should differ by only an imperceptible amount. ...
    (sci.math)
  • Re: Ancient latrine fuels debate at Qumran
    ... You don't get 'truth' about the past from textual analysis. ... We differ on this point. ... Probabilities are all we ever have. ...
    (sci.archaeology)