tedd wrote:
> 
> >  > However, the RND(50) function means that it is returning a number "at
> >>  random" between 0 and 51 (x>0 AND x<51). It may return, and most
> >>  probably will return, the same number more than once in 50 calls. As
> >>  such, then the values of some indexes will be moved to other indexes
> >>  while others aren't moved at all (rather moved to themselves). It's
> >>  an interesting solution and was not obvious to me.
> >>
> >OK, but when a value is not moved, say when index = 1 and the RND
> >function returns 1, this doesn't stop the value at index 1 to be
> >moved later, say when index = 6 and RND(50) returns 1 again.
> >Ingenious, no?
> >--
> >
> >Cheers
> >
> >Alain
> 
> Yes, it is ingenious. However, when it is done that way, then there
> are other indexes which are not touched.
> 
> No disagreement, just investigating the logic.
> 
Tedd,

I know very little about math and I may be wrong but I see it this way:
Ideally, each item will have at one time 49 chances out of 50 to be
moved (98%), plus (counting all the loops) each item has 49 times 1
chance out of 50 to be eligible for a move. Are we not approaching
the certainty? The fact that an item is replaced at its original
location at some time shouldn't count.

-- 

Cheers

Alain

-----------------------------------------------------
FB^3 in Europe:  http://euro.futurebasic.com/
FB II Pouch:     http://www.pixmix.com/FB/outils.html
-----------------------------------------------------