Hmmm CT4 - LOL this was not my first attempt. Found work definately do able. Paper was just LONG really struggled with time and didn't finish. I think there are two type of papers : Extremely tough and able to finish , do able and terribly long. This paper falls into the second. Luckly managed last question however the test was not worth 5 marks... What was with the exposed to risk . was there a trick? Did anyone finish???
Couldn't figure out that exposed to risk question either. I didn't see any difference in the two parts of the question. But I'm fairly sure I'm wrong.
I was really short for time aswell and didn't get it all finished (not my first attempt either unfortunately!) Is the lx symbol in CT4 - I thought it was a symbol from CT5? I couldn't do that question as I couldn't remember how lx was linked to qx or ux
Yeah...lx refers to the number of survivors at age x, and from my knowledge l(x+t)/l(x) = tPx (?...hope Im right here). Was quite sneaky of 'em to put that one in. Sorry to hear that you left it out. I found the Deuce-Advantage question confusing as first...I'm not the fastest when it comes to doing the whole state space and transition matrix thing, and had to scrap my first 2 attempts of that question before writing down a sensible state space and transition matrix (that being said I hope I got 8/8 for those 2 q's cos I left the rest out in that question, not least because I was unsure what do to). The 5-marker in the last question wasn't that difficult in my opinion, but my goodness, it seemed like a 10-marker with the amount of calcs that were required. There 12 were N(ijk)'s and in my understanding we had to calculate the expected value for each of 'em using thier own Bin(n(ij), P(jk)) distributions to calc the Chi-sqaure value. Needless to say, I didn't complete it... Apart from a few annoying theoretical questions, I found the rest manageable. However, I still feel less than 50% sure of passing...a feeling that I'm getting quite used to as I write more of these exams... Goku
Yeah...lx refers to the number of survivors at age x, and from my knowledge l(x+t)/l(x) = tPx (?...hope Im right here). Was quite sneaky of 'em to put that one in. Sorry to hear that you left it out.you correct and l(x+t)/l(x) = exp(-mu(t)) first solve for mu then find 2Px hence l60 - I found the Deuce-Advantage question confusing as first...I'm not the fastest when it comes to doing the whole state space and transition matrix thing, and had to scrap my first 2 attempts of that question before writing down a sensible state space and transition matrix (that being said I hope I got 8/8 for those 2 q's cos I left the rest out in that question, not least because I was unsure what do to).I felt there was two possible transition diagrams - 4 or 5 states The 5-marker in the last question wasn't that difficult in my opinion, but my goodness, it seemed like a 10-marker with the amount of calcs that were required. There 12 were N(ijk)'s and in my understanding we had to calculate the expected value for each of 'em using thier own Bin(n(ij), P(jk)) distributions to calc the Chi-sqaure value. Needless to say, I didn't complete it...Long long long Other question about buns i left out ran out of time- what was the answer for that one q where we were given Pi(3) and had to find transistion prob Apart from a few annoying theoretical questions, I found the rest manageable. However, I still feel less than 50% sure of passing...a feeling that I'm getting quite used to as I write more of these exams...I know what you feeling i get it all the time - hate exams and more i continue more i get discouraged - wrote CA1 aswell wow Goku
I was confused with exposed to risk question - maybe i did it very wrong but was it as easy as just applying trapezium rule
I managed to answer the currant-bun question - not too confident on the methodology and layout of my answer...anyway a mark or 2 from it will be welcome... For the question on the probability of a claim-free year, my final answer was approx 0.76...hehe, a bit spurious hey! Had to solve a quadratic... For Exposed to risk question, yep, for the first part I just used the trapezium method (have a nasty feeling more was required though, if my memory serves me correct, I think that question was 4 marks = 7.2 min, so maybe a bit more was required?)...can't remember what the second part was about, am sure I answered it though (well, hopefully...*gulp!*) Just hope the pass mark is around 57-60%...anything above, and there goes my Xmas-spirit!
I'm not so sure there was a trick with the exposed to risk question. It was only 2 marks for the trapezium rule bit, and then 2 marks for the actuarial estimate for qx. The 4 marks for describing what you'd do with full data was a bit weird, I just described what you would do without having to use the actuarial estimate.
Yeah...I just mentioned stuff along the lines of recording exact date of birth, entry, exit, reason for exit etc. That's how I interpreted that part...good luck though!
I agree with the comment on the paper being do-able but long. But exam pressure is a funny thing. For eg. I do know how to count in tennis, but why oh why did I have to say there is a transition from Deuce to Deuce??I now feel like hitting myself with a tennis raquet (the frame) until there is a90% chance that I will not repeat such silly (silly, cos I should have known better) errors again in an exam. I could not do the buns question properly, either. If I score 1/2 a mark there, I will grab it firmly like Richie Mccaw a turnover ball. As for the NCD question, I got p = 0.5, which is a nice round figure, but it just seems a bit low for the prob. of a claim-free year. That last Cox question, did anybody else remember what they got for the age of the female? I got 65. Goku, about that passing rate...Same here! Eish, I do not want to study this subject again...
Yeah, after 2 attempts I came to a conclusion that the state space was D, A(Andrew), A(Ben), W(Andrew), W(Ben). But that cost me a good 20-25min (!!). Phew! I also got 65! All that effort weren't in vane... Can't say whether p=0.5 for a claim-free year is right/wrong; as long as we showed enough reasonable and logical working, I think we'll both come out with a non-zero mark on that section...good luck man!
yea i got 0.76 as well for my answer for the prob of no claim, what did everyone get for the tennis question, the min no. of points needed for the 90% prob of the game ending? i got 8..
hi got the same space as Goku - 5 states - this is a apparently a standard problem said not irreducible not aperiodic probability of winning the game - i tried to use GP sum, but apparently it is just C-K equations said it was not irreducible, not aperiodic cox - got 65 for female age too. the buns question - i was crossing out attempts but relaised at the end it must be like the general insurance surplus problem with number of buns being a compund poisson process. unfortunately due to time pressure i made it a discrete sum when im sure it was continuous what did everyone get for Km calculation?
Hi Examstudent A compound Poisson Process make sense to me somehow...It is just confusing because there is an increment and increments and Poisson does not go together, normally? I got 50.1 % for the Newspaper Statistic. But I could have make a calc error or two.
I think I also got around 50% for P[T>10] in the KM question...I'm sure it was the last tj interval of my survival function...I took some time on that question working with the time to dismissal...there was one player that had 2 dismissals, but I counted him as one cos we were only interested in the first dismissal...I maybe wrong on this though.
For No claim i got 76%. ended up with a quadratic For Chi squared test i got roughly 7.8 (last one). Didn't finish KM just set it up. Can remember cox questions i answered but i know i started with last two parts of that question and for female i got an age of 65.. Was running out of time. The tennis one was ok i still think it could be 4 or 5 states. 5 states would help answer Andrew winning etc.
goku - KM - i also assumed we were only interested in tiem to first dismissal got 0.5625 but did it in such a rush, im sure calculation errors are present for graduatrion cant remember exact sum, but i summed up sqaured deviations, and concluded it wasnt significant? i.e accept ho
I found the exam alright in general but I think I missed out or fudged far too much overall which will cost me. I only had time for a tiny amount of revision for this and was so pleased as I was going through that I understood what the first 10 questions were talking about. Sadly when I hit 11 with 15 minutes to go I completely lost it and didn't have a clue what it was talking about. Sadly this question was worth 20 marks Can't remember many of my answers, I purged my brain within minutes of walking out the front door, but I did get 8 for the tennis and 65 for the age. The buns one threw me completely though. Starting with K, making c per hour with customers coming in at unknown times... what were we supposed to do there? And were supposed to use the information that "they wanted to minimise wastage by the end of the day" at any point? I spent far too long trying to work out what on earth it wanted here.
i m pretty sure with the buns one your tryiong to model excess buns at any time t so: K + ct - B K =(start buns) ct =(bun production rate * time) B = sum (j =1 to Nt) Bj i.e a compund poisson process as per the surplus in insurance company
exam student - yep, thank Goodness the deviations were given and squared...my accuracy on the calculator was way off before the exam...anyway, from my test, Ho could not be rejected as my chi-square value was less than the critical point... For the currant-bun question, as some of you mentioned, the answer to the second part of that question was a compound Poi. So my answer went like: P(sum(Bi) from i=1 to j > k + ct)...where j is Poi(t(hours))...I'm sure it's wrong as a one-liner seemed insufficient...I took a stab at it nonetheless...
