Course 6-Can someone with All10 help with Stanard's Methods

[cmlarson]

I definitely get the major points of the 4 methods presented by Stanard but now I am trying to get the fine details. All10 has confused me a bit. Here are my questions:

Method 1:
Stanard says in his paper to use a weighted average of the link ratios but I don't see any mention of how many years. Are we to assume that we should use 3 always the way All10 does?

Method 2 & 3:
The calculation of the Expected Losses is supposed to be based on historical loss development experience. In the All10 manual, on exhibits 6 and 7, the Expected Losses are calculated using years 68-71 only. Is there a reason why 1966 and 1967 are ignored in these calculations? Am I missing something?

Method 4:
It looks like All10 in exhibit 8 is using a simple average of the latest 3 year's % of Premium factors. Is that correct? Should any weights be used?

[ramanujan]

I definitely get the major points of the 4 methods presented by Stanard but now I am trying to get the fine details. All10 has confused me a bit. Here are my questions:

Method 1:
Stanard says in his paper to use a weighted average of the link ratios but I don't see any mention of how many years. Are we to assume that we should use 3 always the way All10 does?

Method 2 & 3:
The calculation of the Expected Losses is supposed to be based on historical loss development experience. In the All10 manual, on exhibits 6 and 7, the Expected Losses are calculated using years 68-71 only. Is there a reason why 1966 and 1967 are ignored in these calculations? Am I missing something?

Method 4:
It looks like All10 in exhibit 8 is using a simple average of the latest 3 year's % of Premium factors. Is that correct? Should any weights be used?

Stanard simulates the loss experience for all the years. Exposure does not vary by year. One of stanard's major points is that all years average should be used. If you compare exhibits 6 vs 6a and 7 vs 7a, you will clearly see this difference. Read the comments below the exhibits. Most other authors recommend last three years average.