Actually I heard that they gave credit for 2 answers, after the "solutions"
came out. Someone's appeal was heard.
To: "'Jose Couret'" , studygroup7
cc: (bcc: Jeff Courchene)
From: "Sce, Michael" <MAS2@ChartwellRe.com>
Date: 24.09.98 20:58:36
Subject: RE: Last year's Khury Question
I believe they meant G'(x) and not G"(x). Go by the concave/convex
comments. The question should have thrown out, but it wasn't.
> -----Original Message-----
> From: Jose Couret [SMTP:Jose_Couret@swissre.com]
> Sent: Thursday, September 24, 1998 3:36 PM
> To: studygroup7@lists.casact.org
> Subject: Last year's Khury Question
>
>
>
>
> Number 28 on last year's exam was a "triple true-false" question.
>
> According to the CAS the following part is FALSE.
>
> "3. A concave shaped G(i) (i.e. G"(i)<0) between 0 and n produces a
> more
> conservative confidence interval than does a convex shaped G(i) (i.e.
> G"(i)>0) between 0 and n."
>
> Medical liability for example has G"(i)<0, whereas personal auto
> property
> liability has G"(i)>0. So shouldn't the answer be "TRUE"?
>
> What am I missing?
>
>
>