Using RAND() with Different Distributions!?
 

Hi everyone,

In Excel, by going to Tools/Data Analysis/Random Number Generation,
then clicking ok and going to "Distribution" window, you can select any
distribution available (uniform, normal, ....etc.)

Now, after I make a selection to a distribution, and then using RAND(),
wont that mean that my random numbers do follow the selected
distribution?

Thanks,
Mike

unfortunately no. The rand() function only delivers a random number
that is uniformly distributed between 0 and 1. You can use the
norminv(rand()) function combination to create a normally distributed
random number with a mean and stdev of your choosing.


--
barrfly

Excel User - Energy markets
------------------------------------------------------------------------
barrfly's Profile: http://www.excelforum.com/member.php...fo&userid=4141
View this thread: http://www.excelforum.com/showthread...hreadid=391865

No. the random number generation in data analysis willl generate a table of
random numbers with the given distribution, but it does not change the
Rand() factorr

" wrote:

Hi everyone,

In Excel, by going to Tools/Data Analysis/Random Number Generation,
then clicking ok and going to "Distribution" window, you can select any
distribution available (uniform, normal, ....etc.)

Now, after I make a selection to a distribution, and then using RAND(),
wont that mean that my random numbers do follow the selected
distribution?

Thanks,
Mike

And similarly, you can use other INV functions with RAND() generate
random numbers that follow other distributions. Unfortunately, Pre-2003
INV functions are too crude to do this well (which also means that ATP
does it poorly).

Jerry

barrfly wrote:

unfortunately no. The rand() function only delivers a random number
that is uniformly distributed between 0 and 1. You can use the
norminv(rand()) function combination to create a normally distributed
random number with a mean and stdev of your choosing.

Jerry,
It's worse yet with ATP, because it uses its own rng. It is SO bad that I
would like to know the algorithm. I think I posted some VBA code a while
back to show how lousy it is; you can use it to generate 10000 values; then
sort them and do a frequency count. You'll find some mighty strange stuff;
quite a few numbers get repeated, and some values get repeated 7 or 8 times,
depending on the seed. Suggest to the good folks at MS to eviscerate the
thing, please.

Regards,
Dave B

"Jerry W. Lewis" wrote in message
...
And similarly, you can use other INV functions with RAND() generate random
numbers that follow other distributions. Unfortunately, Pre-2003 INV
functions are too crude to do this well (which also means that ATP does it
poorly).

Jerry

barrfly wrote:

unfortunately no. The rand() function only delivers a random number
that is uniformly distributed between 0 and 1. You can use the
norminv(rand()) function combination to create a normally distributed
random number with a mean and stdev of your choosing.

Hey, David! Good to see you again! How are you?

--
Kind regards,

Niek Otten

Microsoft MVP - Excel

"David J. Braden" wrote in message
...
Jerry,
It's worse yet with ATP, because it uses its own rng. It is SO bad that I
would like to know the algorithm. I think I posted some VBA code a while
back to show how lousy it is; you can use it to generate 10000 values;
then sort them and do a frequency count. You'll find some mighty strange
stuff; quite a few numbers get repeated, and some values get repeated 7 or
8 times, depending on the seed. Suggest to the good folks at MS to
eviscerate the thing, please.

Regards,
Dave B

"Jerry W. Lewis" wrote in message
...
And similarly, you can use other INV functions with RAND() generate
random numbers that follow other distributions. Unfortunately, Pre-2003
INV functions are too crude to do this well (which also means that ATP
does it poorly).

Jerry

barrfly wrote:

unfortunately no. The rand() function only delivers a random number
that is uniformly distributed between 0 and 1. You can use the
norminv(rand()) function combination to create a normally distributed
random number with a mean and stdev of your choosing.

And I repeat Niek's "Hey."

C^2
Conrad Carlberg