I'm set to have a phone interview with GS quant team...Anyone have any ideas what sort of questions will be asked..They said it will be computational, mathematical and problem solving...no finance questions...I've done a search on the forum but if anyone can add any further value it would be much appreciated.

Yeah I got some interviews with them (not over the phone though)- was asked to talk thoroughly about my C++ project which I couldn't remember a single function- some very simple maths and programming questions (integral of ln(x) , write an algo to compute n!, solve a linear diff equation)- some questions about lognormality assumption in BS- a few questions about stoch vol models- a few brainteasers (game theory oriented)- C++ againIndeed no finance question. That's where I realized I really didn't want to be a quant.Good luck!

QuoteOriginally posted by: quantstudent19Indeed no finance question. That's where I realized I really didn't want to be a quant.Good luck!quantstudent: what will you do with an MFE otherwise? assuming that is what you are studying...

QuoteOriginally posted by: energydudequantstudent: what will you do with an MFE otherwise? assuming that is what you are studying...No i'm not doing an MFE (well, I don't think LSE Finance & Economics is considered as such)Anyway, I'm going into trading

Player,Out of curiosity, how did you get your interview? Did you apply online directly to GS? If so, to an associate or to an analyst position?Last year I had a phone interview with FICC ( an FX desk), the questions I got were (a part from some fluffy questions about career plans and the usual):- What is the limit of f(x)=Sqrt{x^2+x}-x as x->Infinity?- If S follows a geometric Brownian motion, what process does 1/S obey? (this is particularly relevant for FX)- One about conditional probabilities with biased coins.- Another one I do not exactly remeber how it went but once you broke it down you just needed to calculate the sum of 2^n with n running from 0 to 64. I solved this by simply using the limit of a geometric series. However, the interviewer mentioned another method involving binary arithmetic (i.e. more computer oriented).I hope this helps and good luck!

1/S is also GBM right? How would you proove this?

Yes it still is a GBM but the parameters change as GBM_(S)[mu,sigma] to GBM_(1/S)[-mu+sigma^2,-sigma] (I am not sure of the sign changes). To prove just use Ito's lemma for 1/S and do the algebra. Afterwards the interviewer mentioned that some people call it Siegel's paradox (or something like that)

i just did it and got:dX/X=( mu - sigma^2)dt + sigma*dWwhere X=1/Sis this correct or have I gone wrong somewhere?

Just did it as well and got the same thing as DAOHow did you get rid of the negative sign?

Sorry made a mistake, my answer didn't make sense. Checked it and got this:dX/X=(sigma^2 - mu)dt - sigma*dWWhich is the same as you guys i think

Yep, got the same here.Another question a friend of mine got at an interview at gs (think this one is rather common, as i had a similar one at another bank)The slope of the zero rate curve is upward. Sort in ascending sequence- the 3 yr zero rate- the yield on a 3y coupon-bearing bond- the fwd rate between 3y and 3.5y

if its upward sloping then the 3 yr coupon rate<3 yr zero rate<fwd rate between 3 and 3.5yDaoApplied by cv ..analyst role...Also I'd be interested in knowing about the 4th question who wrote down....So if you can remember it it would be much appreciated...in addition does anyone know of a brainteaser involving double sided coins? i.e coins which have heads on both sides..I recall some prob question concerning this one but cant quite remember it

Player, The only question I've heard about a double sided coin is the following:Take a bag containing 999 normal coins and one double headed coin. Pick one coin at random and flip it 10 times. If the 10 flips all resulted in heads, what's the probability the coin you chose is double headed? NB1 Maybe the numbers are 'wrong' in this example - I think the answer was supposed to be one half.NB2 Of course, it is just an application of Bayes' Theorem...HTH

Not so sure about the fourth question, but I beleive it went something like this:You have a chess board (a square lattice of 8x8). In the "first" square you put one coin, in the "second" you put two coins, in the "third" you put four coins, and so on... If you fill the whole board like this (the 64 squares), how many coins are on the board?(The answer is obviously the sum of 2^n with n running from 0 to 63, and you can use the limit of a geometric series to know the value of this, though, as I mentioned, there are other methods to solve it)

There are several millions of people on a beach and everyone is looking at someone.What is the probability that no one is looking at you?