Irish Blues

November 1st 2005, 08:07 PM

I answered 31 of 40 questions, *seriously* toyed with marking a 32nd but couldn't convince myself that I had the right answer and it was at the end so I couldn't go back through a 4th time to check the answer.

The statistics part was *hard* - 8 questions at the start, I think I answered 3 of them. From there, a mix of life con and the other stuff. Of the last 22-24 questions, I left 3 blank (the one above was one of the 3, the other will be mentioned below). I feel confident in about 28 of the 31, so I expect that I will pass. However, the stats part may pull the pass mark down a bit.

Also - I'm confident one question *will* be thrown out. It was a question on the difference of the variance between 2 payments, ignore that part. The problem was that we were given "d=6%" and values for 10E50 and 2(10E50). Using these, we needed to find 10p50 - but if d=6%, I got 2 different values for 10p50. It turns out that the problem should have said "i=6%" which then gave 10p50 as .75 - but at that point I knew there was a problem and gave up on it. (Someone else where I took the test thought that even if she used i=6% her answer wasn't one of the 5 listed.)

No Panjer algorithm that I can recall (we're debating this - if it was there, it was *really* simple), no Lee diagram problems (though the concept was helpful for 2 problems), no Poisson-Gamma problems. Several problems with Poisson frequency, some other distribution for severity. The life con had a straight premium problem, a reserve problem, two decrement problems (one needing to find an APV), and two joint-life problems.

The consensus in my room was that we could answer 30 or 31 questions. So throw out the stats and IMO it wasn't a difficult exam. Include the stats and it was much tougher.

The statistics part was *hard* - 8 questions at the start, I think I answered 3 of them. From there, a mix of life con and the other stuff. Of the last 22-24 questions, I left 3 blank (the one above was one of the 3, the other will be mentioned below). I feel confident in about 28 of the 31, so I expect that I will pass. However, the stats part may pull the pass mark down a bit.

Also - I'm confident one question *will* be thrown out. It was a question on the difference of the variance between 2 payments, ignore that part. The problem was that we were given "d=6%" and values for 10E50 and 2(10E50). Using these, we needed to find 10p50 - but if d=6%, I got 2 different values for 10p50. It turns out that the problem should have said "i=6%" which then gave 10p50 as .75 - but at that point I knew there was a problem and gave up on it. (Someone else where I took the test thought that even if she used i=6% her answer wasn't one of the 5 listed.)

No Panjer algorithm that I can recall (we're debating this - if it was there, it was *really* simple), no Lee diagram problems (though the concept was helpful for 2 problems), no Poisson-Gamma problems. Several problems with Poisson frequency, some other distribution for severity. The life con had a straight premium problem, a reserve problem, two decrement problems (one needing to find an APV), and two joint-life problems.

The consensus in my room was that we could answer 30 or 31 questions. So throw out the stats and IMO it wasn't a difficult exam. Include the stats and it was much tougher.