MGFs

[Krieger]

Should we memorize the moment generating functions for all the distributions we'll be encountering?

[krzysio]

You are not going to like the answer, but every time you wonder "should we memorize?", you probably should stop wasting your time on wondering and memorize. One thing I remember from my days of taking exams was that whenever I was looking through the material to study and thinking" "Well ... they can't possibly expect me to memorize this ..." -- that dadgum thing was on the test and required memorization. Look, just memorize it, get it over with, and then you can forget it after the test. Actuarial exams are like a boot camp -- the purpose is not to test your knowledge or your skills -- the purpose is to get rid of you. If you whine instead of studying, you just make it easier for them.

Sorry if this sounds cynical. I am not cynical. I may be wrong in my perspective. But whatever it takes to get you to study, I will do it. There is no way around. The only way is straight through.

Yours,
Krzys' Ostaszewski

[AndrewH]

I wish I had the ability to memorize everything I encountered, but alas I can't memorize everything. I have made a conscious choice that I am not going to waste memory space on the gamma, weibull, and beta distributions. I have elected to memorize the discrete distributions along with the uniform, exponential, and pareto. There's only so much one can put into memory and some can put more than others. However, since I have never seen the gamma, weibull, or beta distribution show up in any of the practice exams I feel it's best to avoid the momorization of these (plus these are the hardest to memorize). Looking at the practice exam P on beanactuary.com I only see the pareto being used twice, and in the context of those problems they were only facilitated by the memorization of the pareto. They weren't that hard to work using first principles =/

[Krieger]

Sounds like a good idea. I'll just look at the distributions in a cursory manner, but I really don't think they'll be on the exam.

[krzysio]

The practice exams posted by the SOA omit significant portions of the material on the test, as they are just past problems from Course 1 since 2000, and there is new probability material on exam P when compared to Course 1. Your approach is flawed, in my opinion.

Yours,
Krzys'

I wish I had the ability to memorize everything I encountered, but alas I can't memorize everything. I have made a conscious choice that I am not going to waste memory space on the gamma, weibull, and beta distributions. I have elected to memorize the discrete distributions along with the uniform, exponential, and pareto. There's only so much one can put into memory and some can put more than others. However, since I have never seen the gamma, weibull, or beta distribution show up in any of the practice exams I feel it's best to avoid the momorization of these (plus these are the hardest to memorize). Looking at the practice exam P on beanactuary.com I only see the pareto being used twice, and in the context of those problems they were only facilitated by the memorization of the pareto. They weren't that hard to work using first principles =/

[AndrewH]

Well if you're not the type of person that can memorize everything then at some point you definitely have to make a decision. Would I like to be able to memorize everything? Well of course! My how much easier that would make exam FM go as well ...

I asked one of the bpp Exam P authors this exact same question a while back and he said he suspected the same thing I did. Of course there's no way to know what's going to be on the exam. The only thing we know is that it's going to be more probability intensive - which is fine in my books.

[managuense]

They no longer provide a density, expected value, moment table?

If they don't...ouch.

[davenpoa]

They no longer provide a density, expected value, moment table?

If they don't...ouch.

They never did supply these for Course 1, only course 3 I think.

[managuense]

Ohh right, I'm an idiot. My version of course 1 was 100/110 and they didn't provide them then. I just got so used to them for course 3 and 4 i got mixed up.

[Arnold Kim]

Honestly, I don't think it's a good idea to just skip out on several different distributions. I've been looking for as much as possible on the new distributions, and yeah it's kind of hard to find info, especially since there are no previous test questions to use as examples, but I think every little bit matters. Will the test be dominated by these problems? Probably not, but I'm pretty sure there will be at least one or two- why mention them on the syllabus if there isn't very good chance they'll be in there? And a lot of the time, one or two questions is what separates passing and failing.

As for mgfs, from my experience (this isn't my first go at Exam 1/Exam P) you could probably get by without memorizing them, but by committing at least a few "basic" ones to memory (like, say, the exponential mgf), you could definitely save some time.