Good luck to all FRM test takers today!

2/5 (you don’t have the problem stated exactly as it was written). The test was very hard, and there were numerous typos! I took it in NY and there were a large number of test takers. I think they drilled very deeply on two or three topics that they kept questioning over and over. I got sick of these questions. How many times do they need to ask a question on X? I loved the probability ones, they were a welcome break. If I passed, it is only because of years of experience in the field. I don’t think you can just read the material and hope to pass. In my exam, the girl next to me got up and left at about 10:30. She never came back. Two others yapped continuously until the test actually started. Another guy was complaining that he felt there was one minute more to go on the morning session and that they cheated him out of it. It’s our responsibility to manage our own time (I learned that painful lesson on the CFA earlier this year) and the proctors DID keep calling out “30 minutes to go”, “10 minutes to go”, “5 minutes to go” and so on. If he had an issue, he needed to speak up before they called “pencils down” instead of expecting them to let him keep writing. Right or wrong, GARP tells you when the exam is over, you don’t. He kept complaining to one proctor after another. No one else had a problem with it. I don’t know if he kept writing after the “pencils down”, I think he did, and that’s why he was making such a fuss. The proctors were very clear about when to return. We started a bit late, and they said to make SURE to be back by 1 pm for the check in. The exam would start promptly at 1:30. After we had all checked back in and had actually received the exam packets and the proctor was reading the rules, four or five people sauntered back from lunch at 1:23 pm! The proctor said that they could not come in. What were they thinking? All that work to prepare, then they can’t take the exam. Anyway, hope everyone did well and can now enjoy a bit of time away from studying. I know I am going to…

That was a trick question. The answer was 2/3. The question stated that there was a 2/3 chance that mail reaches anybody. The probablility of the wife recieving mail doesn’t change just because her husband had a particular outome.

I think the question was not what is the probability that the wife received the letter? If that had been the question, the answer of course is 2/3. But I thought that the question was about the probability that the man didn’t receive a response given that he sent a letter (details are becoming hazy now)… there is a 2/3 probability of the wife receiving the letter, if she receives she replies with certainty, but only 2/3 probability that he’ll get this response, so the answer can’t be 2/3, if I remember correctly. Oh, I just found a link! Check it out. Bayes Theorem, if probability that a letter is lost is 1/n (for us, n = 3) P(X|Y) = (n-1)/(2n-1) so the answer is (3-1)/(2*3-1) = 2/5, this is how I worked it out on the exam, hope it’s right. http://www.actuarialoutpost.com/actuarial_discussion_forum/showthread.php?t=85514

It didn’t ask you about the husband’s probability (which would have simply been 2/3 x 2/3). It asked you about the wife’s probability of having initially received the mail given that her husband didn’t receive anything from her. That would of course be 2/3…her probability couldn’t change. You may be right in the right answer being 2/5 as a lot of questions were poorly written. I wasn’t trying to read too much into what I thought they were trying to ask. I just answered whatever crap question they had on their. Unlike the CFAI, doesn’t GARP eventually release the test questions and their answers?

agree^^. trick question. given that the husband didnt receive any response,they wanted to know the probability of the wife having received the letter. i went with 2/3 after pending 5 minutes on the stupid queston. you can probably find the 2008 questions in the schweser guides next year.

I knew it was an application of Bayes theory, but I don’t memorize formulas and I had a hard time deriving it on the test. I finished 10 minutes early and since I felt pretty sure of the outcome, I decided to spend those 10 minutes trying to solve this question as I knew it was on Bayes updating theorem, but could not set up the problem right. Good job man!

strange! Is it possible that we might have gotten slightly different exams? really, the exam WAS very poorly written. Ok, as a check, anyone remember problem #41 on the afternoon exam, there were numerous typos (they wanted to write “sigma squared” but wrote “sigma with subscripted 2”, they were missing an operator in d2, just said “q sigma subscript 2” not “q - …” and, when they said “what is the likelihood …” the answers had no units, you were just staring at an unadorned number that didn’t make a lot of sense. either way I am so glad this is over!

DoubleDip Wrote: ------------------------------------------------------- > … anyone > remember problem #41 on the afternoon exam, there > were numerous typos … I have absolutely no recollection of anything even close to what you are describing. Doesn’t mean we got different exams though. What would be the best exam strategy we can come up with? For next year’s studying that is…

MehdiOchre Wrote: ------------------------------------------------------- > DoubleDip Wrote: > -------------------------------------------------- > ----- > > … anyone > > remember problem #41 on the afternoon exam, > there > > were numerous typos … > > I have absolutely no recollection of anything even > close to what you are describing. Doesn’t mean we > got different exams though. > > What would be the best exam strategy we can come > up with? For next year’s studying that is… Best strategy in my opinion is to finish your CFA first, then take this test. More than 60% of the material is covered in CFA curriculum, you’ll just have to go over the Basel stuff and the Cerdit and Operational Risks sections.

The “n” in Bayes Theorem is not a whole number but a percentage. As I am not a mathmatician, maybe you could use whole numbers, but I don’t see how…seems like a BS model just like BSM. I thought they were trying to go with Bayes, but given the inputs they gave you, I don’t think it works out.

the question (#41) asked about the probability of a call being in the money, gave formulas for d1 and d2, the typos were in their definitions. yes, had to use black scholes. I don’t agree that you should finish the CFA first then do this. I am taking level II CFA next year. there’s a lot of synergy to exploit, IMHO

Hi guys probability of wife having received letter given man did not receive response is a classic Bayes Theorem question. P (A | B ) = ( P(B|A)*P(A) ) / (P(B|A)*P(A) + P(B|not A)* P(not A) ) where A: probability that wife received letter B: husband got no response such that P(A|B) = ( 1/3 * 2/3 ) / ( 1/3*2/3 + 1*1/3) = 2/5 … I generally found paper 1 very tough. I realized I skipped some sections and did not prepare nearly as much as for CFA. Paper 2 on the other hand was quite easy; yet I do not feel close to confident of passing. Generally had a better idea where I stood with CFA. Anyone remember the quesiton about two two funds that are highly correlated. And which equation is true if you were to combine the two funds into one portfolio. What is true: E(L p) = E(L a) + E(L b) U(L p) = U(L a) + U (L b) RC a + RC b > ULp cant remember the rest. What was the correct answer if you guys know what I am referring to.

markholdt Wrote: ------------------------------------------------------- > Hi guys > > probability of wife having received letter given > man did not receive response is a classic Bayes > Theorem question. > > P (A | B ) = ( P(B|A)*P(A) ) / (P(B|A)*P(A) + > P(B|not A)* P(not A) ) > > where A: probability that wife received letter > B: husband got no response > > such that > > P(A|B) = ( 1/3 * 2/3 ) / ( 1/3*2/3 + 1*1/3) > = 2/5 > > … > > I generally found paper 1 very tough. I realized I > skipped some sections and did not prepare nearly > as much as for CFA. Paper 2 on the other hand was > quite easy; yet I do not feel close to confident > of passing. Generally had a better idea where I > stood with CFA. > > > Anyone remember the quesiton about two two funds > that are highly correlated. And which equation is > true if you were to combine the two funds into one > portfolio. > What is true: > > E(L p) = E(L a) + E(L b) > U(L p) = U(L a) + U (L b) > RC a + RC b > ULp > > cant remember the rest. What was the correct > answer if you guys know what I am referring to. Good job getting that Bayes problem. As for the two funds, it doesn’t matter correlated or not, the total expected loss of the portfolio is the sum of the two expected losses.

If you are right markholdt, then Bayes is utter BS…which is likely true. 2/3 is 2/3 in my book. I’d like to see the mathmatical explanation to why 2/3 is 2/5.

The husband either didn’t receive a letter because the wife didn’t get in the letter in the first place, or the wife got the letter but it didn’t get through to the husband. one might think… Prob (wife’s got it | no response) = Prob ( no response & wife’s got it) / Prob ( no response) = (5/9 *2/3) / (5/9) but wrong since, prob(no response) and prob(wife’s got it) are not independent… Prob (no response & wife’s got it) = 1/3 * 2/3 therefore (1/3 * 2/3 )/(5/9) = 2/5 The answer seems to be 2/5… although I put 2/3 (sigh)

FINforLIL Wrote: ------------------------------------------------------- > If you are right markholdt, then Bayes is utter > BS…which is likely true. 2/3 is 2/3 in my > book. I’d like to see the mathmatical explanation > to why 2/3 is 2/5. I am not sure what you mean by “Bayes is utter Bs” , you’re probably joking, but if you’re not then you might want to read something about Bayesian Statistics. In any case, the 2/3 is the prior probability, you need to update this probability with the new information that the husband did not receive a letter after sending one to his wife, Bayes updating model is used to update existing probabilities based on new information, it’s used everywhere from aerospace engineering to earthquake simulations, …

mo, you definitely got a point ("Best strategy in my opinion is to finish your CFA first, then take this test. ") but I needed something to do right now. Got an e-mail saying that “A record-breaking 13,681 financial professionals took the 2008 FRM Exam … in 78 cities …”

mo, don’t get me wrong, I see where Bayes can be applicable, but from a mathmatical standpoint I would like to see that proven. Particularly in a case such as the question on the exam. You know what the prior probabilty was in that question. Because the relationship is so direct, I need mathmatical proof. I know I can just plug in the numbers, but that isn’t the same as actually understanding the premise.

FINforLIL Wrote: ------------------------------------------------------- > mo, don’t get me wrong, I see where Bayes can be > applicable, but from a mathmatical standpoint I > would like to see that proven. Particularly in a > case such as the question on the exam. You know > what the prior probabilty was in that question. > Because the relationship is so direct, I need > mathmatical proof. I know I can just plug in the > numbers, but that isn’t the same as actually > understanding the premise. I couldn’t solve this question as I said earlier, but the solution posted by markholdt seems correct.

usually Bayes rule is used to make use of additional information to update the apriori probabilities. so the updated probability should be greater than the apriori number…right? because we have more information. 2/5 is < than 2/3. with additional information,the probability has gone down?. sounds weird. maybe i am wrong and there is no such requirement with Bayes