June 10th, 2008, 10:25 am
QuoteOriginally posted by: CuchulainnQuoteOriginally posted by: LapinFDM will be perfect as long as the dimension is small, isn't it?With stochastic correlation models, the dimension increases very quickly (2 underlyings->3 dimensions, 3 underlyings -> 6 dimensions)Of course, for such options yes. I wonder how you would solve a 6 PDE in FDM???? not with the current techniques.I think it's worth stopping for a moment and thinking how good such models are before you solve them.Ok, I've stopped and thought...they are stupid models. I've said this before, and will keep saying it until people start using sensible models: I do expert witness work, and will happily stand up against anyone in court to explain why such models are naive despite their 'sophistication.' They are dangerous. You are going to lose a lot of money if you use them.P