Talk:Credibility theory/Archive 1

Archive 1

DAB

"Credibility" by itself should lead to a diambiguation page with this on it.

Merger proposal

I have suggested that Actuarial credibility be merged with Credibility theory‎ since they are of the same topic: Credibility as actuaries refer to it in the mathematical sense. I think the title Credibility Theory is more appropriate as it tells us more about what it is and not about who uses it. The Actuarial credibility article does have several artifacts that should be preserved however, like the orphan tag, the start-class and the external sources. Anythingapplied (talk) 15:40, 12 January 2009 (UTC)

• Support - it is the same topic. Graeme Bartlett (talk) 20:37, 13 January 2009 (UTC)

I've merged the articles per the above consensus (such as it is) and due to the lack of opposition over the 1,104 days since the last comment to this discussion. I know little about the topic so all I've done is merge it out of the merge backlog. Others can edit it into a unified article. ClaretAsh 10:27, 22 January 2012 (UTC)

Unclarity

This article is very unclear. It seems to me that there are two distinct sorts of policies that could be under discussion here. If j ranges over individuals, then we might have Xj being the amount that individual j has cost the insurance company per year enrolled, while X is the average of all the Xj (that is, the average expenditures per person per year of the insurance company). If j on the other hand ranges over groups, then Xj will be the average amount that individuals in group j have cost the insurance company per year, and X will again be the average of the Xj.

I had assumed that insurers actually operate under a third procedure, which is to take into account all three pieces of information used in the above two procedures (that is, the average cost of the individual to the insurance company, the average cost of members of a particular group, and the average cost of all customers).

Additionally, the "intuitive" clarification of what zj means says absolutely nothing. The credibility of a person is something like how likely they are to lie; the value that is used here is something like the extent to which their past use of insurance is expected to be a good guide to their future use of insurance. In neither case does it mean anything like "acceptability", which is stated as an alternative usage here - I suspect that the author of at least this section of the article is not a native English speaker.

Misleading

The mention of different groups at the beginning of the article seems quite misleading. If motorists are divided by age, sex, and type of car, then there should be at least eight groups. The example contrast of "a young man driving a fast car being considered a high risk, and an old woman driving a small car being considered a low risk" seems to miss the point in a few ways. First, it suggests that the three dimensions of grouping are independent (which may or may not be true), and that risk is monotonically increasing in one direction for each dimension (which probably isn't true for age). But more importantly, it suggests that the companies independently decide which groups are considered high or low risk - instead, the point of this article seems to be to show that the company doesn't have to care which groups have which sort of risk, and can just use the formula mentioned here to calculate an appropriate premium.

Over Complicated

The Actuarial credibility article is much better; it explains the basic concept and gives a simple (and accurate) example. The Credibility theory‎ article tries to use an over complicated example, never accurately explains the concept and uses questionable inferences about homogeneous. Drop the later, rename the former. —Preceding unsigned comment added by 199.244.214.59 (talk) 20:05, 17 June 2010 (UTC)

Expansion

The Credibility Theory article would serve well as an introduction for a credibility article that extends into the different types of credibility and their formulations. Actuarial Credibility article is more of a quick reference guide that can be found on a less informative site than Wikipedia. I would suggest expanding the Credibility Theory article, and removing the Actuarial Credibility theory. — Preceding unsigned comment added by 157.248.5.3 (talk) 14:58, 27 July 2011 (UTC)