[Jprogramming] dyad ": rounding error

[Devon McCormick]

On 3/19/07, Devon McCormick <devonmcc at gmail.com> wrote:
>
> ...
> The advantage of banker's rounding is that it avoids systematic bias.
> >
> >
I take it  back:

   0.5+/ . =1|mm=. mean nn=. ?2 100$100 NB. How many at exactly 1/2?
47
   (+/mm),(+/roundNums mm),(+/(0.5;'U')roundNums mm),+/(0.5;'D')roundNums mm
5099.5 5098 5099.5 5099.5
   0.5+/ . =1|mm=. mean nn=. ?2 100$100 NB. How many at exactly 1/2?
42
   (+/mm),(+/roundNums mm),(+/(0.5;'U')roundNums mm),+/(0.5;'D')roundNums mm
4884 4888 4884 4884

-- 
Devon McCormick, CFA
^me^ at acm.
org is my
preferred e-mail