Actuarial Outpost
 
Go Back   Actuarial Outpost > Exams - Please Limit Discussion to Exam-Related Topics > CAS > CAS Exams > Exam 8 (old Part 9) - Advanced Ratemaking
FlashChat Actuarial Discussion Preliminary Exams CAS/SOA Exams Cyberchat Around the World Suggestions

DW Simpson
Actuarial Jobs

Visit our site for the most up to date jobs for actuaries.

Actuarial Salary Surveys
Property & Casualty, Health, Life, Pension and Non-Tradtional Jobs.

Actuarial Meeting Schedule
Browse this year's meetings and which recruiters will attend.

Contact DW Simpson
Have a question?
Let's talk.
You'll be glad you did.


Reply
 
Thread Tools Search this Thread Display Modes
  #341  
Old 08-13-2019, 02:30 PM
Mud Mud is offline
Member
CAS
 
Join Date: Jan 2018
Location: At home. Occasionally in the basement.
Studying for Exam 8
Favorite beer: Not beer
Posts: 430
Default

Quote:
Originally Posted by trueblade View Post
The Fisher Study Note on the Syllabus? not sure if this helps you

https://www.casact.org/library/study...da_Fisher.xlsx
That doesn't show a graph like I want. I'm trying to literally draw the lines for expected loss and min/max ratable loss, then shade in the insurance charge and insurance savings accordingly. Except, as I change values for those things, the graph and shaded areas should change accordingly.
__________________
ACAS
8 - currently in progress
9
Reply With Quote
  #342  
Old 08-13-2019, 02:41 PM
Mud Mud is offline
Member
CAS
 
Join Date: Jan 2018
Location: At home. Occasionally in the basement.
Studying for Exam 8
Favorite beer: Not beer
Posts: 430
Default

Quote:
Originally Posted by DamSon View Post
They haven't tested a lot of topics covered from Mahler or Robertson either.
Mahler has been *very* lightly tested over the last number of years, and then someone tried to create a problem out of it and the Bailey-Simon paper a few years back that was a disaster [I think Mahler specifically calls it out in his notes and said he wrote a letter saying the question was invalid; the CAS decided otherwise and counted it in full]. I think there's testable stuff in there, it's a question of whether or not anyone writing questions knows how to ask it in an understandable fashion that's sufficiently Bloomsy to pass for an exam question in 2019.

Robertson, on the other hand, ... they appear bound and determined to ask about everything involving now the NCCI went through the process to get to the current hazard group scheme. They haven't asked about the difference between the NCCI approach and the WCIRB approach [yet] or what variables were considered [yet], and I don't think they've touched on the intermediary 4-group system in transitioning to the 7-group system [yet], but they've asked about how the credibility standard was chosen and the different alternatives, the k-means algorithm [including doing a weighted iteration of it], they've asked about Calinski-Harabasz and CCC and calculating the values of each, the "no crossover" principle graphically, and even how underwriters had a say in the final assignment of classes to hazard groups.
__________________
ACAS
8 - currently in progress
9
Reply With Quote
  #343  
Old 08-13-2019, 06:10 PM
DamSon's Avatar
DamSon DamSon is online now
Member
CAS
 
Join Date: Sep 2018
Location: NYC
Studying for 8
Favorite beer: Alcoholic
Posts: 180
Default

Quote:
Originally Posted by Mud View Post
Mahler has been *very* lightly tested over the last number of years, and then someone tried to create a problem out of it and the Bailey-Simon paper a few years back that was a disaster [I think Mahler specifically calls it out in his notes and said he wrote a letter saying the question was invalid; the CAS decided otherwise and counted it in full]. I think there's testable stuff in there, it's a question of whether or not anyone writing questions knows how to ask it in an understandable fashion that's sufficiently Bloomsy to pass for an exam question in 2019.

Robertson, on the other hand, ... they appear bound and determined to ask about everything involving now the NCCI went through the process to get to the current hazard group scheme. They haven't asked about the difference between the NCCI approach and the WCIRB approach [yet] or what variables were considered [yet], and I don't think they've touched on the intermediary 4-group system in transitioning to the 7-group system [yet], but they've asked about how the credibility standard was chosen and the different alternatives, the k-means algorithm [including doing a weighted iteration of it], they've asked about Calinski-Harabasz and CCC and calculating the values of each, the "no crossover" principle graphically, and even how underwriters had a say in the final assignment of classes to hazard groups.
It feels really weird with those two because reading through the papers I thought exam questions on them would be very calculation based with some essay questions, but it seems like the exact opposite. Robertson in particular like you outlined.

I probably need to take a look at past exams to see the question spread, but most of testable portion on section A seems like it just consists of GLM questions.

It seems like knowing GLM, Fisher, Bahnemann, and Clark inside out gets you pretty far on this exam.
Reply With Quote
  #344  
Old 08-13-2019, 06:51 PM
2pacTuary 2pacTuary is offline
Member
CAS
 
Join Date: Oct 2013
Posts: 85
Default

Quote:
Originally Posted by Mud View Post
Mahler has been *very* lightly tested over the last number of years, and then someone tried to create a problem out of it and the Bailey-Simon paper a few years back that was a disaster [I think Mahler specifically calls it out in his notes and said he wrote a letter saying the question was invalid; the CAS decided otherwise and counted it in full]. I think there's testable stuff in there, it's a question of whether or not anyone writing questions knows how to ask it in an understandable fashion that's sufficiently Bloomsy to pass for an exam question in 2019.

Robertson, on the other hand, ... they appear bound and determined to ask about everything involving now the NCCI went through the process to get to the current hazard group scheme. They haven't asked about the difference between the NCCI approach and the WCIRB approach [yet] or what variables were considered [yet], and I don't think they've touched on the intermediary 4-group system in transitioning to the 7-group system [yet], but they've asked about how the credibility standard was chosen and the different alternatives, the k-means algorithm [including doing a weighted iteration of it], they've asked about Calinski-Harabasz and CCC and calculating the values of each, the "no crossover" principle graphically, and even how underwriters had a say in the final assignment of classes to hazard groups.
Do you have a similar write-up for each paper (paragraph 2 above)? To me this insight is gold. If you have it and are willing to share please do so.
Reply With Quote
  #345  
Old 08-14-2019, 10:00 AM
Mud Mud is offline
Member
CAS
 
Join Date: Jan 2018
Location: At home. Occasionally in the basement.
Studying for Exam 8
Favorite beer: Not beer
Posts: 430
Default

Quote:
Originally Posted by 2pacTuary View Post
Do you have a similar write-up for each paper (paragraph 2 above)? To me this insight is gold. If you have it and are willing to share please do so.
Quick gut thoughts:

ASOP 12 - gets touched probably every other year, usually in relation to some other paper. Regardless of how the question is posed, just know you're being asked to regurgitate information.

Bailey-Simon - pretty easy to see what's going to be asked. The question right now is how the writers are going to cloud everything to make it Bloomsy without making it nonsensical. [Thus far, they're doing about 50/50 with that.]

Mahler - covered above. I'll add that stuff in later chapters hasn't been touched in a while, and I'm shocked with each year that no one asks - because it would probably ding half the candidate pool if they did.

Robertson - covered above

Couret-Venter - probably 2/3rds chance to see it. The credibility-weighted relativity piece is a bitch. Just writing it straight-up makes for a perfect exam question. I would make sure to know how to do this. Otherwise, it's going to be a quintiles-based question which should be fairly easy.

GLM - this is getting hit hard. The simple answer is "know it all" but I'll try to refine this as we get closer to the exam.

Bahnemann - this is where ILFs and risk loads live. Know that, especially inflation-adjusted ILFs. Deductibles may get tested, if anything I'm shocked we're not seeing more on disappearing and franchise deductibles. [If we do this time, don't blame me.]

Fisher - I'm throwing everything on retros and experience rating here, including the NCCI RR. The paper is well-written, and this material has long been a favorite for questions. If anything, I think it's not tested enough. Work the case study, do some problems, be comfortable with creating Table M and all its variants.

NCCI ER, ISO ER - feels like it's there and they touch it every other year right now. Not difficult to learn, just requires some time. Spend the time to know it, because it's easy points if you do - and you'll kick yourself if you don't know it and it gets asked.

Clark - there's 197 different questions listed in the paper; aside from perhaps Panjer's recursive method, I think they've all been asked in some form and a few of them will show up. Tends to be relatively heavily tested; expect a couple questions here - especially as it relates to exposure rating.

Bernegger - expect a question off this. I wanted to say "don't worry about solving for parameters for the MBBEFD curves" but it was asked recently, so ...

Grossi-K - also expect a question off this. Exceedance probability is easy, and candidates tend to botch it anyway; eventually they're going to hit the later chapters for obscure, notecard-type stuff that's also easy and which candidates will botch because they glossed it or ignored it completely.
__________________
ACAS
8 - currently in progress
9
Reply With Quote
  #346  
Old 08-14-2019, 10:05 AM
Mud Mud is offline
Member
CAS
 
Join Date: Jan 2018
Location: At home. Occasionally in the basement.
Studying for Exam 8
Favorite beer: Not beer
Posts: 430
Default

Quote:
Originally Posted by DamSon View Post
It seems like knowing GLM, Fisher, Bahnemann, and Clark inside out gets you pretty far on this exam.
I would largely agree with this. In terms of recent importance and difficulty [getting points most don't], I would rank them like this:

GLM
Clark
Fisher

Couret-Venter [tested less, but far more difficult and candidates seem to struggle here]
Bahnemann [tested more, but easier and candidates can scratch out points]

Robertson [less difficult, but seemingly anything is fair game and more likely to put you in an all-or-none situation]
Bernegger [more difficult, but if you can remember a few formulas you can scratch out partial credit]

Bailey-Simon
Mahler
G-K

NCCI ER / NCCI RR / ISO GL

ASOP 12
__________________
ACAS
8 - currently in progress
9
Reply With Quote
  #347  
Old 08-14-2019, 10:12 AM
CuriousGeorge CuriousGeorge is offline
Member
CAS SOA
 
Join Date: Dec 2005
Posts: 1,453
Default

Quote:
Originally Posted by Mud View Post
Bernegger - expect a question off this. I wanted to say "don't worry about solving for parameters for the MBBEFD curves" but it was asked recently, so ...
Reply With Quote
  #348  
Old 08-14-2019, 09:59 PM
CFF CFF is online now
Member
CAS
 
Join Date: Apr 2014
Studying for 789
Posts: 104
Default

In Bernegger source reading page 11, under “calculation of g and b”, how do we actually calculate E*[x^2] in step 2??
Reply With Quote
  #349  
Old 08-15-2019, 02:22 AM
Candidate X's Avatar
Candidate X Candidate X is offline
Member
CAS
 
Join Date: Oct 2015
Posts: 252
Default

Good stuff Mud
Reply With Quote
  #350  
Old 08-15-2019, 10:00 AM
Mud Mud is offline
Member
CAS
 
Join Date: Jan 2018
Location: At home. Occasionally in the basement.
Studying for Exam 8
Favorite beer: Not beer
Posts: 430
Default

Quote:
Originally Posted by CFF View Post
In Bernegger source reading page 11, under “calculation of g and b”, how do we actually calculate E*[x^2] in step 2??
I think it's much more theoretical in nature; I would be surprised if you actually had to do it for the exam. [Asking to recite those steps? That would be incredibly evil and probably on the fringes of what a candidate really needs to know out of this paper as it relates to the exam in total. But, it would also be incredibly CAS-like.]

I'll return to the question of solving for parameters briefly: I think solving for the mean and probability are a given. I think solving for g = 1/p is also totally fair. Solving for b is probably OK given the formulae set in 4.1 on page 10. Stuff like that, ... same comment.

#18, 2017, ... I'm going to admit, I haven't gone through this in detail, but that seems like it's at the upper-end of what they can test as far as solving for parameters. However, I reserve the right to correct this statement as more information comes in.
__________________
ACAS
8 - currently in progress
9
Reply With Quote
Reply

Thread Tools Search this Thread
Search this Thread:

Advanced Search
Display Modes

Posting Rules
You may not post new threads
You may not post replies
You may not post attachments
You may not edit your posts

BB code is On
Smilies are On
[IMG] code is On
HTML code is Off


All times are GMT -4. The time now is 05:21 PM.


Powered by vBulletin®
Copyright ©2000 - 2019, Jelsoft Enterprises Ltd.
*PLEASE NOTE: Posts are not checked for accuracy, and do not
represent the views of the Actuarial Outpost or its sponsors.
Page generated in 0.28649 seconds with 11 queries