I thought it was a pretty tough paper. The questions were either tricky or very time consuming. 16 cells in a chi-squared test was waaaay over the top. Did anyone get the last 7 marks of the compound of a compound question (I think it was qn 10). And did everyone actually derive the expected value/ variance of the compound distribution? I did but for 2 marks I wan't sure if they actually wanted the whole derivation, but as they asked us to derive it, that's what I did. I was pretty pleased that I got qn 2 though, they made something pretty easy look pretty hard.
I assumed it was just the e(s) = e(n)e(x) formula, so I applied that and used it. I noticed on a previous paper that "derive" didn't always mean go into massive detail I think I tried the last compound of a compound one, but it did take me many, many re-reads before I could figure out what they were even asking at that stage. Actually, I initially avoided q10 altogether because they made it sound so complicated, when I don't think it necessarily was. The chi-squared one did seem to go on for a good while... I think I skipped question 5 as I couldn't make head or tail of it - the one with two countries
unexpected questions in CT-3 - April 2013 Dear Mr john Lee it was my second attempt for CT-3 and this time I prepared very well including solving all the previous question papers from 2000 onwards/question bank/assignments etc., and I was very confident this time. But I realized that certain questions were far from routine and some were beyond core reading material . a) Question no – 10 with 6 sub-questions about Compound distribution for 15 marks : The question is about deriving the expressions of mean/variance in terms of certain unknown parameters etc. This question is entirely different from previous diets as this type of question under Compound distribution never came up. If you refer “ Compound Distributions “ core reading material under chapter VII of CT_3, the stuff is just 3 pages and with this stuff it is just impossible to answer this question. Going by previous question papers, there were direct questions on this topic and relatively very easy to attempt. Infact, lot of time had to be wasted in understanding this question as the stuff is entirely new and ultimately it is realized that it is not part of the core reading study material. When I checked up CT-6 study material, I realized that this question can be attempted by students those who prepared for CT-6. This topic is found under Chapter-7 – Risk models-2 of CT-6 from pages 309 to 320. Even this question is there on page no. 315 with little modification. By understanding these 11 pages, one can attempt question no. 10 of CT-3. Those who have prepared for CT-6 only can answer this question promptly. This type of question under compound distribution never cropped up in previous diets. b) Under question no. 5 about finding Confidence interval of poisson process. There is no “n” without which it is difficult to calculate CI and this type of stuff is not there in core reading material. c) Question no. 11 for 21 marks though not a very difficult question but it is not a routine question from chapter Regression, hence consumed lot of time in completing the answer. Most of my friends in exam hall including me were little upset about this. If we do not attempt such a long question, how can we pass. I find that some questions are very damn easy and some of very tough. -Niranjan, India
I agree with the comment below on question 10. I have been going through my CT6 CMP and that question was best fit for a CT6 exam. John, thanks for the exam reviews on youtube !
You did good. The pass rate was 54% which makes it (along with Sept 2011) the lowest in CT3 history...
reply: One or two questions are not part of syllabus, that is the main issue. Naturally, if the questions are not part of syllabus, definitely they look tough.
