SERVING THE QUANTITATIVE FINANCE COMMUNITY

Search found 658 matches

by ThinkDifferent
November 15th, 2007, 1:43 pm
Forum: Student Forum
Topic: Proof of Brownian Motion
Replies: 11
Views: 62766

Proof of Brownian Motion

<r>I mean what I said. continuous paths. Continuous paths + independent&stationary increments => BM. It's a fact, a theorem.please do more research, not just Wikipedia... to save you some time <URL url="http://mason.gmu.edu/~jgentle/csi779/06f/lect12.pdf">http://mason.gmu.edu/~jgentle/csi779/06f...
by ThinkDifferent
November 15th, 2007, 9:01 am
Forum: Student Forum
Topic: Proof of Brownian Motion
Replies: 11
Views: 62766

Proof of Brownian Motion

no, they do not require normality.all you need is 1) continuous paths 2) independent&stationary increments. normality follows from CLT.
by ThinkDifferent
November 14th, 2007, 10:34 am
Forum: Student Forum
Topic: Proof of Brownian Motion
Replies: 11
Views: 62766

Proof of Brownian Motion

don't even need a normality assumption. just show that the process has independent & stationary increments.
by ThinkDifferent
November 6th, 2007, 4:42 pm
Forum: Student Forum
Topic: Stochastic Integral
Replies: 5
Views: 63140

Stochastic Integral

It is normally distributed. This integral can be viewed as a Riemann-Stiltjes integral. just write the integral, by definition, as a limit of Riemann sum. then use the fact that the limit of normals is normal.
by ThinkDifferent
August 17th, 2007, 2:30 pm
Forum: Careers Forum
Topic: Any 8-hours jobs for quants?
Replies: 96
Views: 74429

Any 8-hours jobs for quants?

it does make some sense, cause the waste of your life is compensated with money..
by ThinkDifferent
August 17th, 2007, 1:46 pm
Forum: Careers Forum
Topic: Any 8-hours jobs for quants?
Replies: 96
Views: 74429

Any 8-hours jobs for quants?

<t>QuoteOriginally posted by: katastrofa"I work from 9am to 6pm. If you can do what you're required to do within this timeframe I don't see why would you need to work more."This would be on the verge of acceptability... However, isn't it that when the boss sees that you do all your work in this time...
by ThinkDifferent
August 17th, 2007, 12:30 pm
Forum: Careers Forum
Topic: Any 8-hours jobs for quants?
Replies: 96
Views: 74429

Any 8-hours jobs for quants?

heavily depends on the team. if you are in a front office, I doubt you can get away working <= 8h. I work from 9am to 6pm. If you can do what you're required to do within this timeframe I don't see why would you need to work more.
by ThinkDifferent
August 14th, 2007, 12:21 pm
Forum: General Forum
Topic: lessons for quant investors
Replies: 21
Views: 70536

lessons for quant investors

It is so easy to kick a dead lion...
by ThinkDifferent
July 25th, 2007, 12:08 pm
Forum: General Forum
Topic: Non-homogeneous CTMC simulation.
Replies: 0
Views: 67594

Non-homogeneous CTMC simulation.

<t>couldn't find much related to non-homogeneous continuous time Markov chains... Specifically, assuming we have a CTMC X_t with finitely many states and with time-dependent generator Q(t):suppose, for simplicity Q(t) is a piece-wise constant matrix. So for t in [t_i-1, t_i] we have Q(t) = Q_i - con...
by ThinkDifferent
July 18th, 2007, 2:51 pm
Forum: General Forum
Topic: CIR process; again!!
Replies: 6
Views: 68863

CIR process; again!!

CIR is an affine process, hence a closed form formula is availabe for a Laplace transform. this can be inverted numerically (inverse Laplace transform) to get a density.I am not aware of any analytical results.
by ThinkDifferent
July 5th, 2007, 9:06 am
Forum: General Forum
Topic: local volatility @ Dupire
Replies: 9
Views: 70703

local volatility @ Dupire

<t>As I understand, Dupire formula fails badly if the implied vol surface is a) not smooth b) has arbitrage (you can define what is meant by arbitrage in IV surface). therefore there are other more robust approaches to calculating LV surface. what you are observing may have to do with the above, or ...
by ThinkDifferent
March 23rd, 2007, 9:44 am
Forum: Student Forum
Topic: integral solution
Replies: 4
Views: 75950

integral solution

what's a point of having e^x there if you can factor it out? the indefinite integral e^(t^2/2))dt can't be written in terms of elementary functions.
by ThinkDifferent
March 16th, 2007, 8:33 am
Forum: Book And Research Paper Forum
Topic: how to calculate E(St|ST)?
Replies: 5
Views: 77906

how to calculate E(St|ST)?

just use the fact that E[W_t | W_T=a]=a*t/T (which is not too difficult to prove)
GZIP: On