# 1082 or 1084 for Credibility

Printable View

• November 18th 2013, 12:41 AM
sokura
1082 or 1084 for Credibility
Does anyone else often see people say or use 1084 as the standard for full credibility at auto companies? I was wondering why it seems that a lot of people say 1084 instead of 1082. After all (1.645/0.05)^2 = 1082 where 1.645 is the z-score in your normal table corresponding to the 90% percentile and 0.05 being the tolerable error you accept.
• May 3rd 2018, 07:31 PM
sokura
Found it.
https://www.casact.org/pubs/proceed/proceed68/68175.pdf
It is probably from this ancient paper.
It appears that the algebra along the way to solve for the standard for full credibility, an expansion was used to approximate the standard.
The slight difference it produces is what yields to the 1084.
Maybe actuaries of the past just took this paper's result and started spreading the word to use 1084.
• July 2nd 2018, 07:53 AM
Towanda
Quote:

Originally Posted by sokura
Found it.
https://www.casact.org/pubs/proceed/proceed68/68175.pdf
It is probably from this ancient paper.
It appears that the algebra along the way to solve for the standard for full credibility, an expansion was used to approximate the standard.
The slight difference it produces is what yields to the 1084.
Maybe actuaries of the past just took this paper's result and started spreading the word to use 1084.

Hey you know you are absolutely right here!