in the exam, if we are asked to calculate the variance of a return on a security, based on some data, should we use the unbiased sample variance formula (i.e. n-1 divisor), or the biased one (n as divisor)?
Gareth. I think both will be correct. In such questions I've come accross in CT8 they used biased estimators, while some memo's mentioned that the unbiased estimator is more "correct" but that the unbiased one is used for its simplicity. I reckon you are quite optimistic to think we will get such an easy question! (Seriously, think they will be wanting other levels of understanding in this course) Here is a question for you. I am wondering whether it is beneficial to study chapters 13 and 14 in detail (proofs of 5 step method etc). As nothing like this has ever come up, I reckon it's better to focus on parts of the course that have featured more heavily in past exams and doing extra question practice on that. What's your take on things. Cheers. Erik
it's a very hard call to make, my experience has been whatever strategy i take, the oppisate occurs in the exam, e.g. in ST6 I decided to not to pay too much attention to the old Advanced CID papers, and Murphy's law then put an exact copy of an old ACID question in my exam! I personally have learnt the 5-step method for Binomial model and deriving Black-Scholes, but I think that is more relevent for ST6 than CT8...but it's in the CT8 notes, so they could well be a question on it. As you observed, the CT8/109 past papers seem to focus more on the practical side though... Who knows?
