Actuarial Outpost Fall 2018 LTAM Progress Thread
 Register Blogs Wiki FAQ Calendar Search Today's Posts Mark Forums Read
 FlashChat Actuarial Discussion Preliminary Exams CAS/SOA Exams Cyberchat Around the World Suggestions

 Long-Term Actuarial Math Old Exam MLC Forum

#511
10-27-2018, 06:59 PM
 windows7forever Member SOA Join Date: Apr 2016 Posts: 421

[quote=TranceBrah;9466232]I don't want to roast you too hard so I'll keep my reply short...
- From your comments its obvious you don't have a good grasp on the reserve topic.

QUOTE]

I am not good at solving non-level benefits and non-level expense with other methods, so I did from first principle and prospective method to be safe. That way if I did something wrong in the middle, I would not miss a lot partial credits.

If you did reserve recursion correctly, then you would earn full credits. However, if you had something wrong within any numbers plugged in recursion and just a general recursion formula on the top, it would affect the partial credits in the end.

I just used equivalence principle to calculate EPV(benefits), EPV(expenses), and EPV(premiums) separately, so I get at least full credits for each component I did correctly. That's also easy to check numeric mistakes before lump each part together in the end. They will not keep deducting points from a small mistake in the beginning that get carried over until the end result. That's what I heard from others.
#512
10-27-2018, 07:02 PM
 windows7forever Member SOA Join Date: Apr 2016 Posts: 421

Quote:
 Originally Posted by IacceptTheTerms Exactly what I was thinking! Was debating after the exam with a friend about this haha
Was that something censored and death listed separately? You add censored and death together to get r. Then death (s) / r for each sum together to get H, then e^(-H) = S.

I thought they were going to test something harder.
#513
10-27-2018, 07:08 PM
 IacceptTheTerms Member SOA Join Date: May 2017 Posts: 156

Quote:
 Originally Posted by windows7forever Was that something censored and death listed separately? You add censored and death together to get r. Then death (s) / r for each sum together to get H, then e^(-H) = S. I thought they were going to test something harder.
Yeah that was it but why do you add censored to the risk set though, what's the point of doing this if you already have the risk set at that time? They already subtracted it? Well that's how I understood it and did the same thing as the guy who replied above but I'm not 100% sure, might be wrong I was anxious the whole exam.
__________________
P FM MFE C LTAM VEEs FAP
#514
10-27-2018, 07:22 PM
 IacceptTheTerms Member SOA Join Date: May 2017 Posts: 156

Quote:
 Originally Posted by windows7forever I do not remember the question, but I recalled there was a table with censored and death. I usually draw a line graph to set them up (death, withdrawal, new entry, r)at different times in C, but that question did not need to do these. Or do you think that question was not clear? Or do you think it's death/censored (that censored was r - death - withdrawal already)?
I don't think we're talking about the same question, this one you're talking about is the one that starts with 20 people. I'm talking about the other one where we are given only the risk set at each time and the deaths.
__________________
P FM MFE C LTAM VEEs FAP
#515
10-27-2018, 07:27 PM
 windows7forever Member SOA Join Date: Apr 2016 Posts: 421

Quote:
 Originally Posted by IacceptTheTerms I don't think we're talking about the same question, this one you're talking about is the one that starts with 20 people. I'm talking about the other one where we are given only the risk set at each time and the deaths.
Yeah that's how I got confused. I deleted my comment, since I did not remember the other KM question at all. I worked both out and there's a match with one of choices on both questions. But the one with NA, I was a bit off from the matched choice like a number between two choices but somewhat close to one of choices.
#516
10-27-2018, 07:31 PM
 windows7forever Member SOA Join Date: Apr 2016 Posts: 421

Quote:
 Originally Posted by IacceptTheTerms I don't think we're talking about the same question, this one you're talking about is the one that starts with 20 people. I'm talking about the other one where we are given only the risk set at each time and the deaths.
Did you think retiree and pension MC were easy? I saw each had a majority vote but there were other choices pretty spread in retiree.
#517
10-27-2018, 07:44 PM
 IacceptTheTerms Member SOA Join Date: May 2017 Posts: 156

Quote:
 Originally Posted by windows7forever Did you think retiree and pension MC were easy? I saw each had a majority vote but there were other choices pretty spread in retiree.
Retiree benefits : no idea how to do it but I knew the i=0% so it equals ex-1 I think but couldn't find it. The pension one was a gimme a guess, that is the one to find the replacement ratio?
__________________
P FM MFE C LTAM VEEs FAP

Last edited by IacceptTheTerms; 10-27-2018 at 07:47 PM..
#518
10-27-2018, 08:46 PM
 home_alone Member Join Date: Nov 2008 Posts: 428

Does anyone remember how points were assigned to every WA in the exam?
#519
10-27-2018, 09:24 PM
 windows7forever Member SOA Join Date: Apr 2016 Posts: 421

Quote:
 Originally Posted by IacceptTheTerms Yeah that was it but why do you add censored to the risk set though, what's the point of doing this if you already have the risk set at that time? They already subtracted it? Well that's how I understood it and did the same thing as the guy who replied above but I'm not 100% sure, might be wrong I was anxious the whole exam.
I think I recall more now. The other one you were talking about asked for CI that used Greenwood formula. This one was different. The deaths were very small, and cesnored were large, so the way I said was wrong. That was probably why mine was close to one of five but did not match exactly.

Because deaths were much smaller than censored, do you think both ways will make big differences? In the end number had to be exponented to S, so there should not be a lot difference.
#520
10-27-2018, 09:55 PM
 IacceptTheTerms Member SOA Join Date: May 2017 Posts: 156

Quote:
 Originally Posted by windows7forever I think I recall more now. The other one you were talking about asked for CI that used Greenwood formula. This one was different. The deaths were very small, and cesnored were large, so the way I said was wrong. That was probably why mine was close to one of five but did not match exactly. Because deaths were much smaller than censored, do you think both ways will make big differences? In the end number had to be exponented to S, so there should not be a lot difference.
Depends on what you did? All I did was add up the s/r and exponentiate it? Not sure how many total questions there were on KM/NA, I remeber 2 of them.
__________________
P FM MFE C LTAM VEEs FAP