SERVING THE QUANTITATIVE FINANCE COMMUNITY

Search found 100 matches

  • 1
  • 2
  • 3
  • 4
  • 5
  • 7
by GM
November 2nd, 2009, 5:10 pm
Forum: General Forum
Topic: IR Swap Potential Exposure Calculation with Vasicek model
Replies: 5
Views: 36657

IR Swap Potential Exposure Calculation with Vasicek model

If you get an increasing function rather than a humped curve the problem may be in your PV function. Are you implementing this yourself or calling a pre-defined function? First step though is to check your rate path at the relevant percentile looks sensible.
by GM
June 17th, 2008, 2:54 pm
Forum: Brainteaser Forum
Topic: random walk on a sphere
Replies: 17
Views: 56305

random walk on a sphere

<t>That solution looks correct for the probability of getting to both poles (no absorbing states)But the question asks for the expected time to get to either pole:E1 = 1 + 0.5*E1 + 0.25*E2E2 = 1 + 0.25*E1 + 0.5*E2 + 0.25*E3E3 = E1Solving gives expected time to either pole from the centre E2 = 8Proba...
by GM
May 15th, 2008, 12:41 pm
Forum: Technical Forum
Topic: OTM caps pricing with G2++
Replies: 21
Views: 59467

OTM caps pricing with G2++

<t>Yes, this is a model limitation, it doesn't capture vol smile/skew. However, you should expect a larger relative error in any case, as the option value is smaller for OTM.You can try calibrating to swaptions. The problem with caps is that there is little sensitivity of cap prices to correlation, ...
by GM
April 23rd, 2008, 10:26 am
Forum: Student Forum
Topic: Gamma and dynamic replication - confusion
Replies: 4
Views: 55797

Gamma and dynamic replication - confusion

<t>Hi,I buy a call option and am long delta and long gamma.If I try to replicate the call by dynamically trading the stock, I buy delta units of the stock. I need to own more of the stock as the price goes up and less as it goes down. This is a "buy high, sell low" strategy as described in Hull. The...
by GM
November 23rd, 2007, 12:29 pm
Forum: General Forum
Topic: Help on calibrating the G2++ Model against swaption prices
Replies: 3
Views: 64716

Help on calibrating the G2++ Model against swaption prices

<r>HiGauss-legendre is easy enough to implement, you just need x-values (abscissas) and weights, evaluate the integral at these points, weight them and the answer magically pops out. 2.5s per swaption is too slow for calibration.You can get the weights athttp://<URL url="http://www.efunda.com/math/n...
by GM
October 4th, 2007, 11:42 am
Forum: Numerical Methods Forum
Topic: g2++ swaption formula
Replies: 21
Views: 120519

g2++ swaption formula

<t>A couple of comments about g2++ and swaption calibration:The integral from -inf to +inf is best calculated using Gauss-Legendre polynomial, that will give you a good approx using 10 or 20 points. You can use something simpler like the trapezium rule or brute force but it will be much slower to ge...
by GM
July 26th, 2007, 11:52 am
Forum: Brainteaser Forum
Topic: making number closest to e
Replies: 18
Views: 73105

making number closest to e

8/3 + 5/96 - 2/4170 = 2.71827... is accurate to 4 decimal places... I'm sure there is better.
by GM
July 26th, 2007, 10:20 am
Forum: Programming and Software Forum
Topic: Accessing ranges from a different workbook in VBA
Replies: 2
Views: 67909

Accessing ranges from a different workbook in VBA

Anyone know the syntax needed to do this?My attempt just returns a #value! errorDim searchRange As RangeSet searchRange = Application.Workbooks("G:\MyFolder\MyFileName.xls").Sheets("Sheet1").Range("MyRange")
by GM
January 30th, 2006, 2:28 pm
Forum: Brainteaser Forum
Topic: Switch the rooks
Replies: 9
Views: 122720

Switch the rooks

There is another one similar to bhutes' last solution that also uses underpromotion1. h4 a52. h5 g53. hxg6 (en-passent) Ra64. g7 Rh65. gxh8(R) Rxh1
by GM
January 23rd, 2006, 2:44 pm
Forum: Brainteaser Forum
Topic: Switch the rooks
Replies: 9
Views: 122720

Switch the rooks

1. c4 f52. c5 f43. c6 f34. cxb7 fxg25. bxa8(R) gxh1(R)Can do the same with a and h pawns instead of c and f pawns. Does this count as a solution?!
by GM
October 24th, 2005, 11:14 am
Forum: Careers Forum
Topic: AI and neural networks Phd in a quant career
Replies: 12
Views: 132619

AI and neural networks Phd in a quant career

Nice picture of Antoaneta.Genetic algorithms might be a better topic than neural networks. I don't think there is as much active research in these areas as in the late 90s but I'm not sure why not, it looks like interesting stuff.
by GM
October 18th, 2005, 4:34 pm
Forum: Brainteaser Forum
Topic: Interview Question
Replies: 18
Views: 138652

Interview Question

Ok for a single game it doesn't work, but I'll quote $2.66 as an upper bound for 10 games.
by GM
October 18th, 2005, 4:30 pm
Forum: Brainteaser Forum
Topic: Interview Question
Replies: 18
Views: 138652

Interview Question

I think the distributions for the house and the player should be the same by symmetry. Expected payoff is sumproduct of pdf(house), pdf(player) and the payoff values {1,2,...99}. I'll offer $0.266 for a single game... not sure that is the lowest possible though, it sounds high.
by GM
March 10th, 2004, 5:14 pm
Forum: Careers Forum
Topic: Best UK Universities
Replies: 68
Views: 195137

Best UK Universities

Oxford?? Not for maths or quant finance.Cambridge/Imperial/Warwick would be my top three for undergrad or MSc.
  • 1
  • 2
  • 3
  • 4
  • 5
  • 7
GZIP: On