Kreps Stochastic Model

[UltiActuary]

Originally Posted by UglyChristmasSweater

In my mind, this was just a higher-level question testing your understanding of the safety and variance constraints. (And you have to know what a stochastic model is.)

I thought this may be the case at first two but Section IV is clearly about multiple loss payments and statistical modeling so this clearly was targeting that part of the paper which isn't on the syllabus. Otherwise I could see this being a 'higher order blooms', but it definitely wasn't.

[UglyChristmasSweater]

Originally Posted by UltiActuary

I thought this may be the case at first two but Section IV is clearly about multiple loss payments and statistical modeling so this clearly was targeting that part of the paper which isn't on the syllabus. Otherwise I could see this being a 'higher order blooms', but it definitely wasn't.

We can both be right. It wouldn't surprise me at all if they based the question on section IV but decide to keep the question and accept general answers based on the rest of the curriculum.

[UltiActuary]

Originally Posted by UglyChristmasSweater

We can both be right. It wouldn't surprise me at all if they based the question on section IV but decide to keep the question and accept general answers based on the rest of the curriculum.

Very true, the CAS is mean enough to do that.

[Freudian Slip]

Anyone hear anything on whether they agreed that this question is defective? I can't remember if they let us know ahead of time when they throw a question out.

[David Gao]

I just wrote this was out of the current syllabus in my answersheet and highlighted it as well.

[Freudian Slip]

Anyone received range notices yet? Was part of this question deemed defective?

[Heywood J]

The point total on my fail sheet was 3.75, so I assume no. It doesn't appear that I got any credit for something I thought of on the fly either, which I thought was worthy of something, so I have to assume that that they were looking for the specific answer that wasn't on the syllabus.

[drytellar]

I’m responding here to a set of posts on the validity of the Kreps question from the discussion on waiting for exam results. I think this topic is likely to keep generating a lot of discussion so figured it could be helpful to have it all in one place.

Originally Posted by Heywood J

Assuming the appeal succeeds and this question is thrown out, how does that work? Is the pass mark based on points, or on percentages? If it's based on points, then actually this appeal can only hurt candidates. If it's based on a percentage, then a successful appeal could be a windfall, since it would bump up your grade by about 2.5%, which is a huge chunk of the 5 range.

I’m pretty sure the appeals can’t change the passing score (in points). This actually mostly works to the candidates advantage as otherwise after accepting an alternate solution the Committee could raise the MQC score .

Originally Posted by Heywood J

I have to assume that someone already challenged that question, and I guess it was decided that every competent actuary should know how to answer that question.

As an alternative, I can propose an alternative solution that competent actuaries shouldn't determine risk loads on the fly using methods they don't have training for. Therefore, the valid alternative solution would be to refuse to answer such questions when not given reasonable time. That may fit the appeal window, I hope.[/

This seems like the most likely outcome to me. I do think that a competent actuary should be able to at least BS a half-credit answer to the question. The constraints and strategies are not hard to generalize to multiple periods. We won’t know until sample solutions are released, but I wouldn’t be surprised if the bar for full credit on the question was very low.

I like the sound of an answer that says I won’t determine risk loads on the fly using method I don’t have training for. The exams do seem intended at times to select for the dubious skill of inventing convenient (if wildly inaccurate) inputs on the fly. Accepting that appeal, though, would open quite the can of worms; it could be used as an answer on many questions. It would be like the ASOP inspired equivalent of Cliff Clavin’s final Jeopardy answer: “Who are three people who have never been in my kitchen?”

[Heywood J]

I may not have been 100% serious there. Obviously by that logic almost every answer should be blank to get full credit. I don't think even the actuarial ninja that is Arlie Proctor would prepare answers crucial for their company's financial solvency in 10 minutes, whether he knows perfectly what he's doing or not.

[i++]

If they didn't throw out the part of the question that was off-syllabus... then that is unfair and totally unreasonable IMO. I'm not sure how they handled it, but even grading it with kid gloves or treating it as extra credit would be unjust. It must be thrown out completely.

Actuarial exams are supposed to be testing your knowledge of the syllabus. If they are giving credit for knowing off-syllabus material, then that would mean that there is no possible way for a student to be able to prepare for every question on the test. But if a diligent student studies the entire syllabus and knows it very well, he or she should be able to score 100%. Reading off-syllabus material or coming up with creative answers to off-syllabus questions is beyond the reasonable expectations of a student.