Actuarial Outpost This Irks Me
 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

 Investment / Financial Markets Old Exam MFE Forum

#1
12-06-2010, 05:50 PM
 bigb Member Non-Actuary Join Date: Jul 2010 Posts: 9,175
This Irks Me

Why do they do this? Take for example SOA 69, a question about determining theta under a binomial model.

I can do this problem. This is simply a problem of mechanical calculation. If you have studied the material, this would seem to be a fairly straightforward application of how we determine theta in a binomial model.

In 9 out of 10 problems I have seen, when I get an answer say -0.625, I would expect to see the solution as either -0.625, or -0.63, depending on how they were rounding their answers. I have trained myself that if I'm not exact in my solution, meaning the rounding is off, then I have made an error, and this is a tip to me to go back and be more careful.

Here, there correct answer is -0.6244 something. My point is if you don't carry your intermediate values all the way through, you can easily get -0.625, and like me, since 9 out of 10 times you can round to the number of decimal places the solution gives and you will come to the right solution, here, I would have wasted time going back and second guessing myself because I didn't have exactly -0.625 or -0.63.

Does anyone else feel this is an inconsistency, if so, this question seems to be more of a problem about rounding than it does about applying the material. I don't get what the point is.
#2
12-06-2010, 06:07 PM
 Actuarialsuck Member Join Date: Sep 2007 Posts: 6,155

I'm guessing it's because problems are written by different volunteers and each has his/her preference. With the ability to store values in a calculator why would you round at all?

The only time I have rounded was when using the normal table.
__________________
Quote:
 Originally Posted by Buru Buru i'm not. i do not troll.
#3
12-06-2010, 06:15 PM
 bigb Member Non-Actuary Join Date: Jul 2010 Posts: 9,175

To be honest, in this problem you are pricing an american put, and for me I cant remember every value and where it is stored in my calculator.

I would just like it if some of the little things like that weren't an issue, it is a waste to get stuck on a problem because of the rounding.

I mean, everyone says how accurate and critical actuaries must be in the real world. How the mistake of adding or missing a zero in a calculation dealing with a small number could have a ripple effect and be a serious mistake. All of that is good and dandy, but if these exams are to test my aptitude to see if I can indeed do that in the real world, it would seem to me that actuaries creating the exams should be as critical in developing their questions as they are on the job.

I know your going to say their volunteers and everyone makes mistakes, but it seems a little careless on their end.

Just a gripe, thats all.
#4
12-06-2010, 09:34 PM
 chronoz5707 Member SOA Join Date: May 2009 Studying for FETE(Maybe) College: UTAR Posts: 92

Sometimes accuracy is not always the key to every answer, but the relevance of a value. I can say that there will never be an accurate answer. Certain quarters prefer the use of more decimal points, while others would prefer to use constant number of decimal points (like only 6 decimal places throughout the whole process, for every value obtained).

Just saying that if accuracy is very important, we would not see widespread use of simulation and analysis, but many would waste time solving differential equations. Perhaps we may find an explicit solution of $\int exp{x^2} dx$.
__________________
ASA:
P FM MLC MFE C VEE
FSA:
FETE

Last edited by chronoz5707; 12-06-2010 at 09:43 PM..
#5
12-06-2010, 11:48 PM
 bigb Member Non-Actuary Join Date: Jul 2010 Posts: 9,175

Quote:
 Originally Posted by chronoz5707 Sometimes accuracy is not always the key to every answer, but the relevance of a value. I can say that there will never be an accurate answer. Certain quarters prefer the use of more decimal points, while others would prefer to use constant number of decimal points (like only 6 decimal places throughout the whole process, for every value obtained). Just saying that if accuracy is very important, we would not see widespread use of simulation and analysis, but many would waste time solving differential equations. Perhaps we may find an explicit solution of $\int exp{x^2} dx$.
Please forward your response to the SOA. I'm sure there are a handful of questions over the last three exams I've taken that I've missed due to silly rounding errors.

Thanks
#6
12-06-2010, 11:55 PM
 Actuarialsuck Member Join Date: Sep 2007 Posts: 6,155

Quote:
 Originally Posted by chronoz5707 Sometimes accuracy is not always the key to every answer, but the relevance of a value. I can say that there will never be an accurate answer. Certain quarters prefer the use of more decimal points, while others would prefer to use constant number of decimal points (like only 6 decimal places throughout the whole process, for every value obtained). Just saying that if accuracy is very important, we would not see widespread use of simulation and analysis, but many would waste time solving differential equations. Perhaps we may find an explicit solution of $\int exp{x^2} dx$.
I was pretty certain that it was proven that there is no explicit form for the erf(x) (differential Galois theory?)
__________________
Quote:
 Originally Posted by Buru Buru i'm not. i do not troll.
#7
12-07-2010, 01:00 AM
 chronoz5707 Member SOA Join Date: May 2009 Studying for FETE(Maybe) College: UTAR Posts: 92

Quote:
 Originally Posted by Actuarialsuck I was pretty certain that it was proven that there is no explicit form for the erf(x) (differential Galois theory?)
Forgive my ignorance, perhaps I am still immature.
__________________
ASA:
P FM MLC MFE C VEE
FSA:
FETE
#8
12-07-2010, 07:17 AM
 daaaave David Revelle Join Date: Feb 2006 Posts: 3,103

Quote:
 Originally Posted by Actuarialsuck I was pretty certain that it was proven that there is no explicit form for the erf(x) (differential Galois theory?)
Correct. The proof is pretty cool, but sadly is one of those things I no longer know.
__________________

#9
12-07-2010, 10:56 AM
 ditkaworshipper Official AO Grill Master SOA Join Date: Jul 2009 Location: Ditka's Restaurant Studying for :lol: Favorite beer: Yes Posts: 26,895

Piece of advice: skip American option questions on the exam unless you have a lot of extra time. They take forever.
__________________
Who would win in a fight...Mike Ditka or a hurricane? And da hurricane's name is Ditka.
#10
12-07-2010, 11:26 AM
 rev Member Join Date: May 2009 Posts: 193

Quote:
 Originally Posted by ditkaworshipper Piece of advice: skip American option questions on the exam unless you have a lot of extra time. They take forever.
Then again, since MFE tests a lot on theory, questions with definitive answers are almost like gimmes. I suck with theory, so I saved the American option questions after I put in my guesses for the theory questions .
__________________
P/1 FM/2 MLC MFE/3F C/4 VEEs FAP