Google Groups no longer supports new Usenet posts or subscriptions. Historical content remains viewable.

Dismiss

Probability Distribution in Randbetween

1,131 views

Skip to first unread message

Tom Kosensky

unread,

Feb 6, 2002, 1:45:15 AM2/6/02

This is from an article you can find at:
http://kimura.tau.ac.il/~tali/biometry2000/Biometry_ex02.doc

1. Using Excel or otherwise:
a. Generate N=5 random numbers range between [-1,1] (using function
RandBetween()). RandBetween function in Excel generates numbers with uniform
distribution the mean is ?=0 and the standard deviation ?=0.58.

(I am just posting this information from an article I found - i cannot
verify it's validity)

Tom :)
"max" <demec...@yahoo.com> wrote in message
news:05f001c1aeb0$aec55690$39ef2ecf@TKMSFTNGXA08...
> Hi,
>
> Just a quick check to confirm on the underlying 'default'
> probability distribution used in the Randbetween(x,y)
> worksheet function. Am I correct that Randbetween(x,y)
> uses a uniform probability distribution to generate random
> integers between x & y (inclusive), ie every integer
> within the limits specified stand an equal chance of being
> generated ?
>
> TIA
> Max

max

unread,

Feb 6, 2002, 2:33:51 AM2/6/02

Thanks, Tom. At least there is mention of uniform
distribution therein. cheers. max

>.
>

David J. Braden

unread,

Feb 7, 2002, 9:43:40 PM2/7/02

Max,
That is the *intended* behavior: a discrete uniform. The actual
behavior for versions 4.0 and earlier was wrong: it returned variates
on the lower limit with half the probability of what was returned for
the rest of the support. So if you are working with an old version,
you have to roll your own:
=Lower_bound+Int(rand()*(1+Upper_bound-Lower_bound)

Dave Braden

In article <05f001c1aeb0$aec55690$39ef2ecf@TKMSFTNGXA08>, max
<demec...@yahoo.com> wrote:

> Hi,
>
> Just a quick check to confirm on the underlying 'default'
> probability distribution used in the Randbetween(x,y)
> worksheet function. Am I correct that Randbetween(x,y)
> uses a uniform probability distribution to generate random
> integers between x & y (inclusive), ie every integer
> within the limits specified stand an equal chance of being
> generated ?
>
> TIA
> Max

max

unread,

Feb 8, 2002, 5:23:50 AM2/8/02

Thanks, David. Noted. I'm using XL97 SR2.
So, guess it should be alright. cheers. Max

"David J. Braden" <dbr...@nospam.nokidding.com> wrote in message news:<070220022143408765%dbr...@nospam.nokidding.com>...

0 new messages