Actuarial Outpost
 
Go Back   Actuarial Outpost > Exams - Please Limit Discussion to Exam-Related Topics > SoA/CAS Preliminary Exams > Exam 4/C - Construction and Evaluation of Actuarial Models
FlashChat Actuarial Discussion Preliminary Exams CAS/SOA Exams Cyberchat Around the World Suggestions


Reply
 
Thread Tools Display Modes
  #251  
Old 01-21-2018, 10:27 PM
Beardown9755 Beardown9755 is offline
CAS SOA
 
Join Date: Oct 2015
Posts: 29
Default

3 weeks from E-day!

I've been working through exams from TIA, ADAPT, ASM, and Mahler. I agree with the sentiment from most on the board: Mahler's exams tend to be the most difficult and test the concepts from the syllabus in deeper, more intricate ways than the other sources.

I think the TIA exams thus far (I've done two of them) also do a great job of asking tough questions that require a deeper understanding of the concepts.

ADAPT exams seem great for learning the core of the syllabus, but I feel like they never touch on the rarer topics; they all seem to be very similar in content, while the other sources' exams will have a slightly different feel each time due to diversification of topics.

I'll add my Mahler scores for reference as well:
1) 14/20
2) 12/20
3) 12/20
4) 13/20
11) 29/35
Reply With Quote
  #252  
Old 01-21-2018, 10:36 PM
Beardown9755 Beardown9755 is offline
CAS SOA
 
Join Date: Oct 2015
Posts: 29
Default

Quote:
Originally Posted by llkho View Post
Thank you @dkamka! Now I understood how the variance is 0.75. Although I don't get why we compute the mean by computing 1.5/3.
I work through kernel problems by viewing them as mixtures and proceeding to use properties of mixtures to find CDF, S(x), 1st and 2nd moments, etc., so I don't fully understand the method the SOA used, but it seems to me that they interpret the "kernel" in the solution as a uniform random variable relative to the original data point, that is, the kernel is Uniform(-1, 2), which has a mean of (2+(-1))/2=0.5 and a variance of 3^2/12=0.75.
Reply With Quote
  #253  
Old Yesterday, 10:45 AM
dkamka's Avatar
dkamka dkamka is online now
Member
SOA
 
Join Date: Jul 2010
Location: Austin, Texas
Studying for C
College: Roosevelt University Alumnus '10
Favorite beer: Jawa Juice
Posts: 387
Default Mahler#10

BRUTAL

11/20 mucho malo. That was tough as it gets so far. I think #10 is the hardest from Mahler I've seen yet.

This is good because learning from these is key to getting better.

Starting full length Mahler tonight. Looking forward to it because there may be a better chance of passing with more questions. Each exam title page says that "each question is similar to a question in the study guide". This is why I regret using ASM/TIA. If I used Mahler all along I'd be golden by now. These would be cake because they're all the same and they'd be like repeat questions. maybe not cake, but not as bad. I'm feeling Mahler puts me at game day level 100% of the time where other resources are not always there.
__________________
PM me if you’re in the Austin area
Exams: P | FM | MLC |MFE | C
VEE: Finance | Economics | Applied Statistics | Accounting

Predictive Analytics
FAP
APC

Last edited by dkamka; Yesterday at 03:21 PM..
Reply With Quote
  #254  
Old Yesterday, 03:31 PM
dkamka's Avatar
dkamka dkamka is online now
Member
SOA
 
Join Date: Jul 2010
Location: Austin, Texas
Studying for C
College: Roosevelt University Alumnus '10
Favorite beer: Jawa Juice
Posts: 387
Default

Quote:
Originally Posted by Beardown9755 View Post
I work through kernel problems by viewing them as mixtures and proceeding to use properties of mixtures to find CDF, S(x), 1st and 2nd moments, etc., so I don't fully understand the method the SOA used, but it seems to me that they interpret the "kernel" in the solution as a uniform random variable relative to the original data point, that is, the kernel is Uniform(-1, 2), which has a mean of (2+(-1))/2=0.5 and a variance of 3^2/12=0.75.
From my last Mahler exam (#10), I got almost an identical question. This is how I understand it for uniform:

Just calculate the mean from the constraints:
E[X]=E[(y-1+y+2)/2]=E[(2y+1)/2]=Ey+1/2. --add the boundaries of y-1 & y+2 and divide by 2 as uniforms do --> (a+b)/2. When you have the Ey, that means use the empirical value from the data, so 5 and add the 1/2. This is why it says it's mean is NOT the same as the empirical estimate. The variance for each data point is uniform distribution with bandwidth of 3. So the way we calculate variance is empirical plus the uniform or (3)^2/12 + 6 where 6 you get from calculating it on the calculator and finding sigma squared. Does that make more sense?
__________________
PM me if you’re in the Austin area
Exams: P | FM | MLC |MFE | C
VEE: Finance | Economics | Applied Statistics | Accounting

Predictive Analytics
FAP
APC

Last edited by dkamka; Yesterday at 03:33 PM.. Reason: typo correction
Reply With Quote
  #255  
Old Yesterday, 04:17 PM
Beardown9755 Beardown9755 is offline
CAS SOA
 
Join Date: Oct 2015
Posts: 29
Default

Quote:
Originally Posted by dkamka View Post

Just calculate the mean from the constraints:
E[X]=E[(y-1+y+2)/2]=E[(2y+1)/2]=Ey+1/2. --add the boundaries of y-1 & y+2 and divide by 2 as uniforms do --> (a+b)/2. When you have the Ey, that means use the empirical value from the data, so 5 and add the 1/2. This is why it says it's mean is NOT the same as the empirical estimate. The variance for each data point is uniform distribution with bandwidth of 3. So the way we calculate variance is empirical plus the uniform or (3)^2/12 + 6 where 6 you get from calculating it on the calculator and finding sigma squared. Does that make more sense?
Yeah, that makes sense.

Also, the first full-length Mahler exam was a little easier than the shorter ones IMO, although it was still a tough exam.
Reply With Quote
  #256  
Old Yesterday, 04:40 PM
Adapt and Chill Adapt and Chill is online now
SOA AAA
 
Join Date: Sep 2017
College: Davidson College
Posts: 8
Default

Quote:
Originally Posted by dkamka View Post

Starting full length Mahler tonight. Looking forward to it because there may be a better chance of passing with more questions. Each exam title page says that "each question is similar to a question in the study guide". This is why I regret using ASM/TIA. If I used Mahler all along I'd be golden by now. These would be cake because they're all the same and they'd be like repeat questions. maybe not cake, but not as bad. I'm feeling Mahler puts me at game day level 100% of the time where other resources are not always there.
Isn't that the point of the Mahler exams though? You're seeing a lot of questions for the first time that you don't immediately know how to do. It gives you more exposure to a variety of problem types under test-taking scenarios.

Is this your 2nd sitting? I'm using ASM and a co-worker is using TIA, and am curious about the difference between the two.
Reply With Quote
  #257  
Old Yesterday, 05:41 PM
ngiletto ngiletto is offline
CAS SOA
 
Join Date: Oct 2017
College: Alumni - George Washington University
Posts: 29
Default

Quote:
Originally Posted by Adapt and Chill View Post
Isn't that the point of the Mahler exams though? You're seeing a lot of questions for the first time that you don't immediately know how to do. It gives you more exposure to a variety of problem types under test-taking scenarios.

Is this your 2nd sitting? I'm using ASM and a co-worker is using TIA, and am curious about the difference between the two.
This is not my first sitting, and I did not feel TIA prepared me for the twists and turns on the exams, but prepared me for the base concepts. I suggest Mahler, as I am trying it myself this sitting. So far, it does seem to have more questions that throw me for the loops the exam threw me for.
__________________
P FM C
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 06:05 AM.


Powered by vBulletin®
Copyright ©2000 - 2018, 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 0.26852 seconds with 10 queries