Re: binocdf inaccuracy in Octave

[Daniel J Sebald]

On 07/08/2013 04:22 PM, Daniel J Sebald wrote:

> So, we should never really be doing any such kind of
>
> ret = 1 - numlibfunc (...)
>
> However, do a search
>
> [sebald@ scripts]$ grep " 1 -" * -r
>
> and the scripts, at least, are littered with all sorts of operations.
>
> Rik, you really discovered something here...looks like you have your
> work cut out for you! :-)

I suppose this is really a cosmetic thing.  I mean, on the other end

> > k = 1
k =  1
> > betainc (1-1e-3, k+1, 50-k)
ans =  1

is losing it's accuracy as well, i.e., it should really be 1 - 1e-150.

Dan