<t>QuoteThe choice of the penalty parameter is tricky.Yes, this is the key issue and the one I had not implemented correctly. It is neatly answered in Forsyth on p.8 where they give a guidance of 'tol' = 1 / 'Large', where 'tol' is used to terminate the application of the penalty, and 'Large' is the...

<t>Thanks, Chuchulainn,We need to factor in stability into those inequalities ;-)I nailed the issue in my code. I had rho (the penalty parameter) set to 100,000 which is certaintly "Large" but it looks as though it over-penalises the solution for breaking the LCP formulation. Irony (for me) is that ...

Oh, thanks. Odd, I'm sure it was giving different values earlier. Well C=0.18 is in line with Call-Put Parity. I beginning to believe:a) My penalty method is buggy or sensitive to something.b) Haug's (and other's) implementations of BAW or other routines are also buggy.

Great, thanks. Can you please post your Java BAW code?

<r>Somewhere between .15something and .16. 4 out of 6 online pricers give me a .15ish value. My Forsyth implementation gives values .15something for the most 'expensive' calculation (actually just below .16 for some method parameters).A few examples:<URL url="http://www.varcalc.com.au/calculators/va...

<t>Hello Cuchulainn,Thanks for your reply. What is your justification for saying the 0.13 value is correct? Barone-Adesi & Whaley is also wrong in that case.QUADRATIC CONVERGENCE FOR VALUING AMERICAN OPTIONS USING A PENALTY METHOD by P.A.Forsyth and K.R.Vetzhal, 2002, SIAM Journal of Scientific ...

<t>Hi Folks,I'm working on a project and coming across something interesting which I'm having a hard time finding out about. I found that under a purely Explicit FDM or Crank-Nicolson FDM then I get a price of around 0.136 for an American put with S=K=T=1, vol=40%, r=0.05%, but using a Penalty Metho...

QuoteOriginally posted by: DCHHi Aankz,This was written in 1997 - I wonder if anyone has written papers on this subject recently...Sorry, I meant 2007, and by "this" I mean Schoutens article, not Moodley's thesis.

<t>Hi Aankz,Sorry for the late reply - I wasn't following this thread earlier in my life :-)I agree with you that the "-i" is a mistake and should be "-1". It's not consistent with Moodley's formula in the same paper, nor with Carr and Madan's paper, nor with Simpson's Rule itself. Pleoni et al, in ...

Forgive me, Alan. What you say is an absolute load of correctness! It comes out in about 6 or 7 lines.You would think authors would put occasional hints in about how expressions come about, or perhaps I'm just more stupid than everyone else ;-)Thanks,David.

Thanks for the reply, Alan.I think this is in the right direction, but I'm not sure your integral gives what we want. int(z,infty) e^-y = e^-z so we lose any integral of e^-y which appears on the RHS of Musiela's expression.

<t>Hi Folks,I'm reading the section on Lookback options from Martingale Methods in Financial Modelling by Marek Musiela and Marek Rutkowski, and have got stuck on one expectation:This is on p. 240 of my copy (2nd Ed, 3rd Printing). I just can't see how this works - can anyone explain, step by step, ...

<t>Just to follow up. I have found a lot of Q&A's. Google and pdf search engines and a lot of time/patience are all you need. Just start with terms like "stochastic calculus exercises".There is no published book full of Q&A's although Oksendahl's book has lots of questions and a few worked s...

Yes - I found the errata after posting (sorry). Yes, it's the same.

Thanks. I realised dt fits with Girsanov too as we integrate.