OASIS Mailing List ArchivesView the OASIS mailing list archive below
or browse/search using MarkMail.

office-comment message

[Date Prev] | [Thread Prev] | [Thread Next] | [Date Next] -- [Date Index] | [Thread Index] | [List Home]

Subject: Re: [office-comment] ODFF: CRITBINOM constraints on Alpha

• From: Regina Henschel <rb.henschel@t-online.de>
• To: office-comment@lists.oasis-open.org
• Date: Wed, 31 Aug 2016 18:57:47 +0200

Hi Eike,

Eike Rathke schrieb:
Hi,

On Wednesday, 2016-08-31 14:02:29 +0200, Eike Rathke wrote:

6.18.19 CRITBINOM
in Constraints says "Alpha >= 1", that makes no sense and instead should be
"0 <= Alpha <= 1"

Or better "0 < Alpha < 1" instead? Excel and Calc do not allow values
0.0 and 1.0, but Gnumeric does and returns 0 for Alpha=0 and number of
trials for Alpha=1.

Opinions?

I think Gnumeric calculates it correctly.

CRITBINOM returns the smallest value k with P(X<=k) >= Alpha.

For Alpha=0 the return value is always 0, because a probability is always >= 0 and therefore already P(X=0) fulfills >= Alpha=0.

For Alpha=1 it returns n besides in case p=0. For p<>0 you need to sum up all P(X=k) from k=0 to k=n to reach 1. Only in case p=0, already P(X=0)=1 >=Alpha=1 and therefore the return value is 0. [If an outcome is impossible, then the event X=0 is always true.]

The results for the cases Alpha=0 and Alpha=1 are mathematically well defined. So there is no need to exclude them.

That is my personal opinion.

Kind regards
Regina

[Date Prev] | [Thread Prev] | [Thread Next] | [Date Next] -- [Date Index] | [Thread Index] | [List Home]