What type of Quant do you want to be? You could do a masters/Phd. that deals with statistics/econometrics which will be useful in a lot of hedge funds and would make you an attractive hire to them if you Phd. is in a useful area.

Doesn't sound too promising for your career goals IMO...

<t>QuoteOriginally posted by: DoubleTroubleQuoteOriginally posted by: ACDAre they from a multivariate normal distribution or are thy just marginally normally distributed?They are not from a multivariate normal distribution. I have X_1, ..., X_n random variables which are all N(0,1).Ok can't really h...

Are they from a multivariate normal distribution or are thy just marginally normally distributed?

<t>These pointers I think may help. x(t) is distributed as a normal variance-mean mixture. Normal Variance Mean mixtureIf you could observe z(t) this would be fairly trivial to estimate. Depending on the distribution of z(t) it may be fairly easy to determine the distribution of z(t) | e(t) and then...

<t>QuoteOriginally posted by: martingullHiIsnt it just integration by parts. where f(x) = tau, f'(x) = 1, g(x) = B(tau), g'(x) = dB(tau)so that its tau * B(tau) - B(tau)Or am I misunderstanding you?Is that right? I don't use Ito calculus much (so forgive me if I'm talking crap) but I would have thou...

Thanks for the pointer, Hansi's third picture looks fine to me. A download link would be fine too (anything that avoids having to type the code out of the book).

Any chance of this being made avilable on the Kindle (or is it already? I got one recently and really like the thing)? Would it even work in that format? I assume there's a cd with the book, do amazon allow for CD bundling with a Kindle purchase? Thx

Or another way (think this should work):[diag(s)]^(-1) C [diag(s)]^(-1)Where the diag operator takes a column vector and returns a diagonal matrix with the vectors elements along the diagonal.

Also you could standardise them by their in sample volatility so that their movements are of similar magnitudes (if you don't care about the volatilities).

<t>QuoteOriginally posted by: outrunThat was all I was saying:"it's impossible for the sum of observed data to have a negative variance"*but* you're moving to a different discussion: "the origin of faulty correlation matrices"There are lots of way to handle missing data when estimation correlations!...

<t>QuoteOriginally posted by: outrunI'm talking about actual data: e.g x1={3,1,4,1,5,2,6,5,...}, you can construct a linear combination of vectors that give a vector who's elements have negative variance. You are talking about transform of properties of variables, I'm talking about actual values of ...

<t>QuoteOriginally posted by: outrunQuoteOriginally posted by: ACDBecause it is invalid in this situation. If it wasn't (semi) positive definite then you could construct a linear combination of the variables in the matrix whose sum has a negative variance.I think there is subtle difference: its more...

Because it is invalid in this situation. If it wasn't (semi) positive definite then you could construct a linear combination of the variables in the matrix whose sum has a negative variance.

Neither term refers to a Barrier option in my experience, they just mean that the payoff is the Heaviside step function (i.e. a fixed amount or nothing, hence the names as there are only 2 possible outcomes) and are inter-changeable.