stottlemeyer - If I were in your position, I would do the Part III, then consider doing a PhD. This will leave open the option of following an academic career in the future.

Ditto. Enjoy your free time.

For the iTraxx index, for example, you have the 0-3, 3-6, 6-9... etc. tranche prices.But can anyone tell me what the quote for the "index" represents?Does it represent the running premium payable on a CDS on a single name of the basket?Thanks.

An interesting topic is optimal investment under credit risk.

<t>Zedr0n,If you take X to be a discrete random variable on 3 states, say, then this problem is an instance of the problem I stated in my first post, where x, y, and z are the probabilities corresponding to each state.If X is continuous, then yes, you could take the Frechet derivative of F or G with...

Ok. Then my original problem involving expecations (see my 2nd post) is not solvable, as this assumption does not hold for the probability mass function or density function.Thanks for your help rcohen.

<t>Yes, but my thinking is that you have complete freedom to choose the endpoint (x1, y1, z1). This endpoint is not uniquely determined by dG in general. To illustrate:F(x, y) = 2x + 3yG(x, y) = 8x + 9y.dF/dG = 2 + 3(dy/dx) / 8 + 9(dy/dx).But what is dy/dx ?? Implictly differentiating gives:dF/dx = ...

<t>rcohen, thank you for your replies.The reason why I still think the question is ill-defined is that a small change dG does not necessarily result in a unique dF. This is because the perturbation dG could be linked to dx, dy, or dz in many (infinitely) different ways, each way resulting in a (poss...

<t>rcohen, you said:This I understand, but how do you calculate the total derivatives dy/dx and dz/dx ??It would seem that since x, y and z are independent variables, the derivatives must be zero.Actually, what got me thinking about this question is the following:What is dF/dG where:F = E[ f(X) ]G =...

I have 2 smooth scalar valued functionsF = f(x, y, z) G = g(x, y, z).What is dF/dG ??

If you have a bivariate student t distribution P(X < x, Y < y), is the conditional distribution P(X < x | Y = y) a student t distribution?

Could someone explain how to calibrate a jump diffusion model of stock prices to historical stock price data?

<t>I'm currently a PhD student in 1st year. I have an offer for an internship at a relatively small investment bank, in trading. I would like to complete an internship during my degree, but was hoping to do one at a big bank, and also preferably in a quant role rather than trading. And I can't do mo...

<t>Dates available for internship: end of June --- end of SeptemberLocation: I am in the UK, but am willing to work anywhere in the world.Status: PhD Financial Maths in top uni, BSc (Hons) Mathematics (1st class).Age: 24.Preparation: Various maths courses, including Stochastic Calculus, PDE's, Numer...

<t>QuoteOriginally posted by: quannabePersonally I'd choose Applied Math or Statistics. At where I am (a top 5 American university) students from these departments have no difficulty being placed in the finance industry.The demand is strong for Finance PhDs but they tend to be get research type jobs...