<t>Hi Cuchlain/spv205QuoteCould you post how you discretised exactly mixed terms?For the ADE scheme, I have done:a) create a full matrix operator with two-sided differences (ie, matrix M)b) then virtually wrote M=U+L for two triagonal matrices... and solved the problem as proposed in the ADE scheme....

<t>Hi CuchulainnQuoteNo dt/2 terms in ADEUps, that's a typo. Will correct it.QuoteI use Barakat and Clark (B&C) which averages the U and V schemes and is 100% explicitActually, I have implemented that, too, but it seemed to work less well (not scientific argument, just observation).[QI think I k...

<t>Hi CuchulainnUpdate on this - I have finished the implementation including the forward PDE for both SVLV and stochastic dividends. I have implemented ADE with parallel splitting and with step-wise splitting (first X then Y). I have compared this with ADI. Basically, I have found for my particular...

You are mixing up risk-neutral and real-life expectations.

<r>I am looking (again <E>;-)</E>) at forward-PDE calibration of stochastic vol-local vol.The aim of the exercise is to see whether this beats the ADI scheme we have already implemented.DetailsMy PDE is a forward PDE for exp OU stochastiv vol and an equity with an additional local vol term. The func...

I heard somebody calling that "Rumourtrage"

Has somebody used the ADE schemes Duffy mentions for the 2F case with mixed derivatives and drift?

Hi DileepThis thread is ages old but in case you are still interested, here is a new paper on truly stochastic dividends (not only affine).Any comments very welcome

<t>I think traden4alpha hit the point.QuoteThe safest conclusion is that a rapidly growing economy makes person A's SOL more volatile and puts pressure on person A to boost their productivity or to adapt to changing economic conditions (e.g., learning to weave fish-farm enclosures or saving money to...

<t>Most commonly used is to writew_new = w_old + w_pos - w_negie, split your transaction into positive and negative part. You constrain both of them to be positive. Then your transaction cost model isc'w_pos + c'w_neg + rav/2 var( w_old + w_pos - w_neg ) = (*) this gives you a semi-definite covarian...

<t>If you have no collateral agreement and you recevie no recovery upon default, then you can use:Your asset is S(t), and default of the counter party has no effect on it.r(t) = risk free rated(t) = div yieldk(t) = instantaneus credit riskThe you'll have to pricein other words you change discounting...

<t>You could try the model on pg. 36 - "Fitted Heston" which is also described here on page 48ff with some comments on practical implementation and hedging performance.The basic idea is to use a given SV model, and fit it to the prevailing variance swap prices. The resulting model has a very low num...

<t>Hi cchoongYour question is not precise enough. What do you need your correlation for?Assuming you want to hedge your European option on a basket, then the correlation which drives your PnL is the correlation between your hedging instruments. In theory, that should be the correlation of the forwar...

Which tools are you referring to?Are you saying that your PHP written constrained LSQ optimizer beats professional optimizers such as Mozek's ?

<r>It's not the forward because NDFs are usually traded against currencies which are not freely convertible ... CNY for example.For the forward price argument to work, you need to be sure that you can exchange the two currencies in the future. Hence there is conversion risk. As a first approximation...