Actuarial Outpost
 
Go Back   Actuarial Outpost > Exams - Please Limit Discussion to Exam-Related Topics > SoA/CAS Preliminary Exams > Exam 3/MLC - Actuarial Models
FlashChat Actuarial Discussion Preliminary Exams CAS/SOA Exams Cyberchat Around the World Suggestions

Actuarial Jobs by State

New York  New Jersey  Connecticut  Massachusetts 
California  Florida  Texas  Illinois  Colorado


Reply
 
Thread Tools Display Modes
  #451  
Old Yesterday, 07:06 PM
Jim Daniel's Avatar
Jim Daniel Jim Daniel is offline
Member
SOA
 
Join Date: Jan 2002
Location: Davis, CA
College: Wabash College B.A. 1962, Stanford Ph.D. 1965
Posts: 2,558
Default

Quote:
Originally Posted by PerpetualMotion View Post
Can someone ELI5 the differences in these variances from Fall 2014

(ii) Drug A will be given to 1000 subjects comprising Cohort A.
With probability 80%, each subject in Cohort A will have q = 0.20 .
With probability 20%, each subject in Cohort A will have q = 0.05 .
Conditional on q, the lives have independent future lifetimes.

** I understand that everyone in the group will experience whichever q is chosen so we use (un?)conditional variance formula. I don't really know why this is though.

(iii) Drug B will be given to 1000 subjects with independent future lifetimes,
comprising Cohort B. The drug will affect different subjects differently
and independently.
With probability 80%, any given subject in Cohort B will have
q = 0.20 .
With probability 20%, any given subject in Cohort B will have
q = 0.05 .

** This one you end up finding E(Q) and conditional variance of Q because any given subject can experience different Qs.


I really just don't understand this topic, so if anyone can break it down I'd really appreciate it. Or if someone can point me to TIA videos or other online resources. This is a major flaw in my understanding of variance and I'm really trying to understand.
ii) says there is an 80% probability that you have 1000 trials with q = 0.2 [so a Binomial RV with m=1000 and q = 0.2] and a 20% probability that you have 1000 trials with q = 0.05 [so a Binomial RV with m=1000 and q = 0.05]. Thus the distribution is an 80/20 mixture of these two Binomial RV's. Get the mean and second moment by mixing and then get the variance. [My website has a free study note on mixture distributions.]

iii) says that each individual has an independent 80% probability of its q being 0.2 and a 20% probability of its q being 0.05. Thus, by mixing, each individual has a probability 0.8 (0.2) + (0.2) (0.05) =0.17 of death, and there are 1000 such individuals, so the number of deaths is Binomial with parameters 1000 and 0.17. So the mean and variance just come from this Binomial.
__________________
Jim Daniel
Jim Daniel's Actuarial Seminars
www.actuarialseminars.com
jimdaniel@actuarialseminars.com
Reply With Quote
  #452  
Old Yesterday, 07:15 PM
PerpetualMotion's Avatar
PerpetualMotion PerpetualMotion is offline
Member
SOA
 
Join Date: Jan 2013
Location: Kentucky(Louisville)
Studying for Exam MLC
College: University of Louisville Alumni
Favorite beer: Zombie Dust
Posts: 325
Default

Wow, that actually makes perfect sense. I'm going to check out that study note. I really appreciate it, I just know there's going to be a question like this on the exam
__________________
P,FM MFE , C, MLC
Modules 1-8 IA
FA


And NUH is the letter I use to spell Nutches
Who live in small caves, known as Nitches, for hutches,
These Nutches have troubles, the biggest of which is
the fact there are many more Nutches than Nitches.
Reply With Quote
  #453  
Old Yesterday, 07:37 PM
Steveo1794 Steveo1794 is offline
Member
SOA
 
Join Date: May 2016
Location: Ohio
Studying for MLC
College: Miami University (OH)
Favorite beer: Anything
Posts: 32
Default

Quote:
Originally Posted by PerpetualMotion View Post
Wow, that actually makes perfect sense. I'm going to check out that study note. I really appreciate it, I just know there's going to be a question like this on the exam
Did this question last night... Confused the hell out of me but once I really thought about it, it made sense.

I think about it like this:
The scenario in b(ii) has five total groups of 1000 each. In four of those five groups, everyone has q=.20. In the other one, everyone has q=.05.
The scenario in b(iii) is one group of 1000. 80% of the 1000 have q=.20 and the other 20% has q=.05.

Hope this helps.


---

Side note - Are we allowed to use pens on the written answer? Obviously not on the MC, but the WA?

Last edited by Steveo1794; Yesterday at 07:40 PM..
Reply With Quote
  #454  
Old Yesterday, 07:42 PM
Jim Daniel's Avatar
Jim Daniel Jim Daniel is offline
Member
SOA
 
Join Date: Jan 2002
Location: Davis, CA
College: Wabash College B.A. 1962, Stanford Ph.D. 1965
Posts: 2,558
Default

Both of Steveo1794's interpretations are incorrect.
__________________
Jim Daniel
Jim Daniel's Actuarial Seminars
www.actuarialseminars.com
jimdaniel@actuarialseminars.com
Reply With Quote
  #455  
Old Yesterday, 08:21 PM
Steveo1794 Steveo1794 is offline
Member
SOA
 
Join Date: May 2016
Location: Ohio
Studying for MLC
College: Miami University (OH)
Favorite beer: Anything
Posts: 32
Default

I wasn't referring to the literal scenarios presented in the problem, I was referring to how to distinguish between the "each subject will have" and "any given subject". That phrasing confused me, and thinking about it like that helps me.

Does the below explanation make sense? If not, please correct it.

Scenario b(ii) is 80% chance that 100% of the cohort will have q=.20 and a 20% chance that 100% of the cohort will have q=.05.
Scenario b(iii) is simply an 80% chance of a member of the cohort having q=.20 and a 20% chance of someone in the same cohort having q=.05.
__________________
P FM MFE C MLC
VEE FAP APC

Reply With Quote
  #456  
Old Yesterday, 08:45 PM
Jim Daniel's Avatar
Jim Daniel Jim Daniel is offline
Member
SOA
 
Join Date: Jan 2002
Location: Davis, CA
College: Wabash College B.A. 1962, Stanford Ph.D. 1965
Posts: 2,558
Default

Quote:
Originally Posted by Steveo1794 View Post
I wasn't referring to the literal scenarios presented in the problem, I was referring to how to distinguish between the "each subject will have" and "any given subject". That phrasing confused me, and thinking about it like that helps me.

Does the below explanation make sense? If not, please correct it.

Scenario b(ii) is 80% chance that 100% of the cohort will have q=.20 and a 20% chance that 100% of the cohort will have q=.05.
Scenario b(iii) is simply an 80% chance of a member of the cohort having q=.20 and a 20% chance of someone in the same cohort having q=.05.
My problem with how you described it is the following: correct solutions based on the statements you gave yield incorrect answers for both cases. However, you may have interpreted your statements in an incorrect way that produced correct answers in boith cases. It's good to be lucky.

Your explanation of b(ii) is correct.

Your explanation of b(iii) may be correct, but the wording strikes me as a little confusing. It would be clearer to say: For each member of the cohort, there is an 80% chance that that member's q is 0.2 and a 20% chance that that member's q is 0.05, and the results are independent from member to member.

By the way, I agree that the exam problem as stated is poorly worded, with the "any given subject" and "each subject" difficult to distinguish---and imagine what it's like for non-native English speakers! Of course, exam questions are often poorly worded, and it's not all that rare for them to use textbook terms to mean something different from their meaning in the textbook, or to be missing data, or for the wrong answer to be considered correct. Quality control truly sucks. Candidates need to be robust problem solvers in that the first part of the problem is to figure out what the problem really is.
__________________
Jim Daniel
Jim Daniel's Actuarial Seminars
www.actuarialseminars.com
jimdaniel@actuarialseminars.com
Reply With Quote
  #457  
Old Yesterday, 09:44 PM
MySpaceTom MySpaceTom is offline
SOA
 
Join Date: Oct 2016
Posts: 7
Default

The Written Answer instructions at the beginning of the exam state " Do not write answers to more than one question per sheet".

Does this mean every part (a, b, and c) to question 1, must be on an individual sheet? Or, is it okay to answer parts a, b, and c all on the same sheet?
Reply With Quote
  #458  
Old Yesterday, 10:38 PM
PerpetualMotion's Avatar
PerpetualMotion PerpetualMotion is offline
Member
SOA
 
Join Date: Jan 2013
Location: Kentucky(Louisville)
Studying for Exam MLC
College: University of Louisville Alumni
Favorite beer: Zombie Dust
Posts: 325
Default

Quote:
Originally Posted by Jim Daniel View Post

By the way, I agree that the exam problem as stated is poorly worded, with the "any given subject" and "each subject" difficult to distinguish---and imagine what it's like for non-native English speakers! Of course, exam questions are often poorly worded, and it's not all that rare for them to use textbook terms to mean something different from their meaning in the textbook, or to be missing data, or for the wrong answer to be considered correct. Quality control truly sucks. Candidates need to be robust problem solvers in that the first part of the problem is to figure out what the problem really is.
I'm the first to admit that life isn't fair but that really seems like bs from the soa standpoint. We put in hundreds of hours worth of work for each of these exams, they should be held to a higher standard. Some problems are unnecessarily complex and it just doesn't feel right. The further I progress, the more it all seems like a cash grab. I hope when I'm an FSA I get to write problems that actually tests relevant knowledge. Until then.
__________________
P,FM MFE , C, MLC
Modules 1-8 IA
FA


And NUH is the letter I use to spell Nutches
Who live in small caves, known as Nitches, for hutches,
These Nutches have troubles, the biggest of which is
the fact there are many more Nutches than Nitches.
Reply With Quote
  #459  
Old Today, 12:10 AM
NoFunnyStuff NoFunnyStuff is offline
Member
SOA
 
Join Date: Jul 2009
Location: Green Bay, WI
Studying for MLC
College: Georgia State University Alumni
Posts: 164
Default

Quote:
Originally Posted by noonelikesmodules View Post
Doing a profit/gain question. Specifies surrender and mortality RATES. Doesn't specify about timing of surrenders. Do we always assume EoY surrenders?
Most of the time yes, because if you have already paid for your insurance in BOY you would not surrender before the end of the year for something you have already paid for.
Hope that this help.
__________________
Prelim:P FM C MFE MLC
VEE: FINANCE ECONOMICS STATISTICS
Reply With Quote
Reply

Thread Tools
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 08:06 AM.


Powered by vBulletin®
Copyright ©2000 - 2017, 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 0.18662 seconds with 10 queries