Replace =RAND() with =ABS(RAND())

"Bernard V Liengme" wrote in message
...
Has anyone experimented with RANDBETWEEN? With RAND() one needs to do many
many recals to get neg values. But RANDBETWEEN(0,5) gave me lots of
negative
values the first time I made a sheet with the formula repeated 500 times.


I presume and hope this refers to Excel 2003 since I am still using the
previous version even if the results of RAND() are not as well distributed
as they might be. There is no indication of trouble with RANDBETWEEN() in
that version.

I think Microsoft needs to take quick action like Intel did with the Pentium
bug many years ago. AFAIK, there were no reported financial losses due to
the Pentium bug but that might not be the case with RAND(). The ambulance
chasers and class actioners must be licking their lips!


--
James V. Silverton
Potomac, Maryland, USA

"James Silverton" wrote...
...
I think Microsoft needs to take quick action like Intel did with the Pentium
bug many years ago. AFAIK, there were no reported financial losses due to
the Pentium bug but that might not be the case with RAND(). The ambulance
chasers and class actioners must be licking their lips!

We can only hope.

--
Never attach files.
Snip unnecessary quoted text.
Never multipost (though crossposting is usually OK).
Don't change subject lines because it corrupts Google newsgroup archives.

Hi Bernard,
I didn't spot the problem earlier, but I was waiting for MS to specify the
algorithm before doing more than perfunctory testing. It has been a long
wait g. To your question, the seed for Rand() is apparently set by the
clock. Were you to restart your experiment, you would get a different
result. I do, for su under Win XP, I can't yet replicate what you found,
though I don't doubt it. I'm still plugging away on a fast machine.

Regards,
Dave B

"Bernard V Liengme" wrote in message
...
Has anyone experimented with RANDBETWEEN? With RAND() one needs to do many
many recals to get neg values. But RANDBETWEEN(0,5) gave me lots of
negative
values the first time I made a sheet with the formula repeated 500 times.

Also noticed another oddity with RAND(). With a block 2000 by 12 of
=RAND()
formulas and a macro to repeatedly do recals, I found that when negative
values occur there are always 16.0 to 17.2% of them.

Bernard

"Patrick O'Beirne" wrote in message
om...
What a failure - imagine all the simulation models affected, never
mind the office lottery syndicates!

If anyone has stories of failures from this, the European Spreadsheet
Risks Interest Group would like to hear of them:

http://groups.yahoo.com/group/eusprig

www.eusprig.org

Patrick O'Beirne

"Harlan Grove" wrote in message
news
"James Silverton" wrote...
..
I think Microsoft needs to take quick action like Intel did with the
Pentium
bug many years ago. AFAIK, there were no reported financial losses due to
the Pentium bug but that might not be the case with RAND(). The ambulance
chasers and class actioners must be licking their lips!

We can only hope.



As I recall in the case of the Pentium bug, Andy Grove sent a letter to me
personally apologizing and offering prompt free replacement (machine
produced probably, but still!). I wonder if we see this from Bill Gates!


--
James V. Silverton
Potomac, Maryland, USA

"James Silverton" wrote...
...
As I recall in the case of the Pentium bug, Andy Grove sent a letter to me
personally apologizing and offering prompt free replacement (machine
produced probably, but still!). I wonder if we see this from Bill Gates!

I doubt it. More likely he'll baracade himself behind the disclaimers in the
EULAs which basically say that anyone who uses any Microsoft software has no one
but himself/herself to blame if that software does anything, either desired or
undesired. If you read the EULAs rather than the marketing BS, Microsoft has
never claimed its software does anything useful.

--
Never attach files.
Snip unnecessary quoted text.
Never multipost (though crossposting is usually OK).
Don't change subject lines because it corrupts Google newsgroup archives.