[libreoffice-users] Re: Missing function: Bankers Rounding

Hi Regina,

Am 19.02.2011 15:16, Regina Henschel wrote:

> I think, MROUND will work for Robert. Examples:
>
> Round to nearest even Cent:
> MROUND(12.354;0.02) results in 12.36
But the bankster rounds to 12.35

> MROUND(-12.328;0.02) results in -12.32
But the bankster rounds to -12.33

> Round-to-even to one decimal place
> MROUND(3.49;0.2) results in 3.4
But the bankster rounds to 3.49

> MROUND(2.51;0.2) results in 2.6
But the bankster rounds to 2.51


With "normal" rounding, 1/1000 digits ...
1,2,3,4 round down to previous zero
5,6,7,8,9 round up to next zero
This means that there are 20% more ups than downs because of the 5 which 
is exactly in the middle but rounds up by mere convention.
The oh-so-clever ones round up or down depending on the second last 
digit (1/100) being even or odd when the 1/1000 digit is 5. So the 5 
goes up in 50% of all cases and down in the other 50%.
2.515 => 2.51 [5/1000 down]
2.525 => 2.53 [5/1000 up]
This can be solved by some carefully designed script or sheet formula 
but I don't feel penuriously enough to do so.

http://www.tech-archive.net/Archive/Excel/microsoft.public.excel.misc/2004-04/5994.html
gives strange results or I do not understand it. It is generic enough to 
be run by this interpreter.


--
Unsubscribe instructions: E-mail to users+help@libreoffice.org
List archive: http://listarchives.libreoffice.org/www/users/
*** All posts to this list are publicly archived for eternity ***