trying to prove out negbinomdist

[Eike Rathke]

Hi Nicholas,

On Wednesday, 2020-06-03 08:49:01 -0400, Nicholas Ferguson wrote:

> This might be an unusual question for this site.  But I cannot prove out the negative binomial distribution, by so called hard coding the function per LibreOffice docs….a jpg of that function is attached.  Does anyone have a clue. Though I will add…Microsoft Excel has the same problem.   I did search for its corresponding code in C++/H in libreoffice source…but that didn't help either…
>  
> =negbinomdist(1;14.4;.92)
> 
> =(FACT(1+14.4-1)/(FACT(1)*FACT(14.4-1)))*(power(.92;14.4)*power(1-.92;1))

You are overlooking that X and R are defined to be integer values, so
=negbinomdist(1;14.4;.92)
effectively is calculated as
=negbinomdist(1;14;.92)
and if you substitute 14.4 with 14 in your manual approach to form
=(FACT(1+14-1)/(FACT(1)*FACT(14-1)))*(POWER(0.92;14)*POWER(1-0.92;1))
you'll discover that both results are almost equal, apart from some
precision error of 5.55111512312578E-17

See also the ODF Formula (ODFF) definition at
https://docs.oasis-open.org/office/OpenDocument/v1.3/cs01/part4-formula/OpenDocument-v1.3-cs01-part4-formula.html#NEGBINOMDIST

  Eike

-- 
GPG key 0x6A6CD5B765632D3A - 2265 D7F3 A7B0 95CC 3918  630B 6A6C D5B7 6563 2D3A
-------------- next part --------------
A non-text attachment was scrubbed...
Name: signature.asc
Type: application/pgp-signature
Size: 833 bytes
Desc: not available
URL: <https://lists.freedesktop.org/archives/libreoffice/attachments/20200603/44125341/attachment.sig>