OK - my so-called "proof" is obviously off a little bit, as shown by my last
example.  This is why I like simulations.

On Wed, Sep 3, 2008 at 5:04 PM, Devon McCormick <[EMAIL PROTECTED]> wrote:

> ...
> If we look at the distribution for my original problem of five things, we
> see
>
>    mm=. rpm 1e6$5
>    +/mm =/ i.6
> 365976 375745 166694 83220 0 8365
>
> This zero near the end is because there can't be only four matches, so this
> is why this result doesn't agree with the theoretical percentages in the
> paper mentioned by Henry.
>
...
-- 
Devon McCormick, CFA
^me^ at acm.
org is my
preferred e-mail
----------------------------------------------------------------------
For information about J forums see http://www.jsoftware.com/forums.htm

Reply via email to