Anyone have any thoughts on this April's CT3? I couldn't manage question 6 i think it was with T=XY with Y/X=x ~ N(x,1) Anyone feel the same?
That question just clean bowled me. Anyway, the paper was vvvvvvvvvvvvvvvvvvvvvvvvvvvvvv time pressured. Overall, a paper of that standard should be given 4 hours. The examiner gave me a real battering in this paper. Maybe next time..........................
Question 6 Yeah I remember that. We had to make two simulations of T=XY, where X~N(200,10?) and Y|x ~N(x,1) I made two simulations of X first... using the standard normal tables at the back. If you look up a number from U(0,1) on the cumulative probabilities you can find the corresponding standard normal z value and then convert to N(200,10). Then I plugged in the two simulations of X into Y|x to get two new normal distributions that I could simulate in the same way (using the remaining two U(0,1) numbers given). T=XY can be simulated by multiplying the simulations of X and the corresponding Y|x. The question took me 20 mins to do... it was worth 6 marks... that was a bummer.
Don't remember the question number but I didn't appreciate the one where you had to sketch the likelihood function of the uniform distribution and give the MLE of theta Plus one of the longer questions (about investor satisfaction in a particular type of fund) took forever. Did seem like a hard paper!
not again... I found the paper quite tricky too (for the 3rd time!), but I don't think it was any harder than the past papers. They all get you to apply principles given in the notes in slightly new ways, which can throw you if you don't have a complete inside-out and back-to-front knowledge of the material (which I didn't). I think I was relying on being able to figure it out in the exam (which usually works), but the time pressure means you haven't got any time to think. I think the secret for this subject is practice practice practice. Fingers crossed, they'll set the pass mark at 50 and I'll sneak it! Otherwise, 4th time lucky. Oh, and I spent far too long crunching numbers on that stupid contingency table question about satisfaction with investments. Grrr.
Yeah I didn't like that theta one either, and it came right after the T=XY one, not a nice couple of questions, 11 marks down the pan
Don't mean to be rude. But 4th time for CT3?? Is this one a stumbling block. Have you got many others?
Same here. Infact very tricky. Exactly my feelings. Lets hope its 40/100. A wicked question, far to much time was required on it that it was worth. After all it involved a histogram, a Chi2 table twice, the calculations, theory, etc, etc,
I only took 2 papers this sitting, but both of them I was still writing after 3 hours! Crazy. It's usually nice to have a few minutes to check back over your work as I always make a standard level of stupid mistakes Really annoyed with myself though as I got through the first 9 questions in an hour and a half and thought that I was going fine and relaxed a bit...then took an hour and a half on the last 3 and didn't finish the last part of the last question And those last questions were worth crazy numbers of marks! I have to admit to sitting staring at the N = XY question for about 5 minutes before it clicked as well. Barmy question...but it made you think Glad it's over for the moment though! Leafy
Hmmm Hi there folks - well seems the consensus is tight on time and the Q6 generate normal random variates was grotty, along with the sketch of the uniform likelihood function in Q7 and the stacks of calculation needed for the (unused) contingency table in Q10. Fair point - CT3 always has about one 'odd' question to test your ability to apply your knowledge (remember skip this and come back to it later - get all the 'standard' question marks first!). For Q6 SpY was right - but it would've thrown a lot of people . I'll stick something in the notes for next session on this - would people like this in the FAQ on our website as well? The uniform likelihood one is a bit weird - you'll find it on page 10 of Chapter 10 as a special example (as the "differentiate it and set it equal to zero" method doesn't work). If there's anything that we could add to the notes that might help future students do let me know either in the forum or email me at john lee.
