How was the time-eating monster? Did anyone manage to finish in time. I left about 20 marks worth of questions, rest was ok; but this paper was definitely harder than the other attempts. Any comments?
I learned my lesson on CT5 Start with the last questions as they give more marks per time spent on the exam. It's the fault of the institute that the exams are like this but if you want to pass you might as well "play the game". I too did not finish but only left four marks worth of questions.
I thought the exam was fair. For the question on MLE the part when it told us the find the variance of the CRLB for the unbiased estimators, what were the two estimators? I know one of them was just the variance of 1/n x sum of x. What was the other one. Think there were four marks going on this so i know there must be another one.
Hi guys! I am suddenly panicking because I can't remember that part of the question. I remember deriving the MLE and CRLB and showing the confidence interval at the end...what else was there? Did I just miss out a whole part of that question? I found the paper mostly ok (assuming I actually answered all the questions, which I thought I had until 3 minutes ago ), but I didn't have time to check over any work at the end. I was too busy making up my own rules for covariances, which of course turned out to be all wrong, so dropped 5 marks right there. I was quite happy not to see any nasty conditional expectation questions (that I was aware of, anyway...?)
As I recall the CRLB was for only one estimator, but I could be wrong. The number of marks quite often seems to be not well related to the amount of work required. Could anyone do the correlation coefficient of a linear transformation question with only a small amount of work. I don't understand how that was worth only two marks.
I think the worst I have seen is in CT1. They regularly ask for an explanation of the no arbitrage assumption. The standard answer is that there is no risk free profit. Somehow or another this is worth two marks and not one. In the recent CT1 a question was split up into two parts both worth 3 marks. Naturally one part was twice as long as the other.
test question One of the longer questions asked you to test the means then test the variances of two samples. I found that the means were the same and also the samples had a common variance? This seemed unusal to me. How did everyone manage with this?
That is what i got as well. I think what the question was trying to test is that from the dot plot that you drew you would have seen that the spread of one of the samples was different from the other. So by first asking us to test whether the means were the same (which i found that they were the same i.e. i got a probabiltyvalue of around 60%) then asking us to test the variance of the samples to see if they were the same given that the dot plot showed that they were different is to see whether the conclusions in the first t test would be valid or not i.e. ifwe found that the variances were different then we would have concluded that the result of the t test was also invalid. I think i got a probabilty value of around 10% for the variance test. I can't remember if i multiplied this by two given that it was a two tail test. Did anyone else get similar values as me.
April 2007 I remember seeing the pass mark for this was 82%! Tbh, I thought this paper was one of those "silly" papers with fiddly questions (Albeit with small marks). I think it was more tough than the others. I had 45 - 1 hour spare when I did it, no foolin'. I tend to go steam rolling through. I thought the Central Limit Question part (iii) [poisson bit] was vague and couldn't figure out I needed to find P(x > 4). MLEs I always find hit n miss. Some are tough, some are straight forward. This was so-so.
