"James Silverton" wrote:
Joe wrote on Tue, 9 Mar 2010 09:54:20 -0800:
"James Silverton" wrote:
Technically, if you do not allow duplication you don't truly have random
numbers.
[....]
Wrong! There is random selection "with replacement" and
"without replacement".

My aside was somewhat a matter of math definition and meeting
tests for randomness in a population of numbers. I am not "wrong".

LOL! Okay, more precisely, random selection "without replacement" means
random selection from different populations, each population being the same
as the previous ones less the previously selected things. There is nothing
that is inherently "not truly random" about it.

Of course, whether or not the selection is "truly random" depends on the
selection process. No software pseudorandom generator (AFAIK) is "truly
random". They only try to achieve the appearance of randomness.

But that has nothing to do with whether or not duplication is allowed, which
is all that I was commenting on.


----- original message -----

"James Silverton" wrote in message
...
Joe wrote on Tue, 9 Mar 2010 09:54:20 -0800:

"James Silverton" wrote:
Technically, if you do not allow duplication you don't truly have random
numbers.

So by your definition, it is impossible to choose 6 people "at
random" from a group of 49.

Wrong! There is random selection "with replacement" and
"without replacement".

----- original message -----

"James Silverton" wrote in
message ...
mike_vr wrote on Tue, 9 Mar 2010 09:19:01 -0800:

Does anyone know how I can have six cells next to each
other, using the =RANDBETWEEN(1,49) function, where none of the cells
equal each other? I.e. a random lottery function
without having two of the same numbers. I'm trying to do
this without creating a huge nested If statement so any help
would be appreciated.

Technically, if you do not allow duplication you don't truly have random
numbers. However, there are sites that will
produce both duplicated and unduplicated "random" digits, for
example, http://stattrek.com/Tables/Random.aspx These could be put in a
column and selected
sequentially.

So by your definition, it is impossible to choose 6 people "at random"
from a group of 49.

Wrong! There is random selection "with replacement" and "without
replacement".

My aside was somewhat a matter of math definition and meeting tests for
randomness in a population of numbers. I am not "wrong".

However, for all practical purposes any method will do. Almost all
generated "random" numbers are pseudorandom anyway. I admit that there are
hardware solutions using electronic noise but even they will produce
repeats. It's analogous to the fact that, say, 10 zeroes will occur
sequentially in the digits of PI.

--

James Silverton
Potomac, Maryland

Email, with obvious alterations: not.jim.silverton.at.verizon.not