You can always perform a Monte Carlo simulation of the local vol model. I think that's the standard way of testing if a model's dynamics are realistic.

Try using the local vol model to price a forward starting option, invert the price to get the implied Black vol surface. Repeat for different strikes, expiries, forward starting points. You will see why the predicted future vol surface is not believable.

Unless I am mistaken, the question is relating to writing options on variance (i.e. at a particular maturity and variance strike), rather than just swaps. This is quite exotic and you would hard pressed to find a trading desk that readily gives you two-way prices on these.

<t>This is pretty standard. You first determine your Heston vega, i.e. the change in PV by bumping each single Heston param in turn and re-evaluating your trade. Through your calibration closed form formula you should also have a matrix relating the Heston params to the market implied vol surface. W...

From trade date, add 3 months, then minus 2 business days.

For this case, your firm's history is a good teacher. And typically cubic spline interpolation is needed to reduce the effect of the kink at the point where the futures meet the swaps. After all is done, always plot the implied forward curve to check if anything seems badly wrong.

You also experience the same problem with barrier options if the underlying gets near the barrier. I had 'some fun' with exactly this situation involving a large barrier option trade last week.

You don't really have a reasonable semi-static hedge for these, that's why traders typically charge a fat margin related to 'pin risk' before they take on these trades. At least that's what I do.

Wrong! An FRA is a martingale only under its associated T-forward measure. As for why futures are martingales under the risk-neutral measure, think about margining and try figure out its implication.

At my place, correl sensitivity is termed just correlation vega, which is probably less likely to be misunderstood compared some other more fancy names.

That's right, in other words, pdf is proportional to the second derivative of call price with respect to strike. But you only get terminal distributions, nothing about the conditionals from european options.

Sorry to spoil the party but you are thinking of pricing spark spreads this way then you will be basically screwed. You have to model the spread directly, forget about modeling the underlyings and infering the spread. Won't work. Ask any seasoned pros in the area.

Thanks, was thinking along somewhat similar lines. I guess we can't run away from introducing a few simplying heuristics on cross factor correlations.

<t>Let's say I am pricing a cross asset hybrid product which is dependent on the CMS spread of a particular currency and a commodity with a futures curve that has a significant probability of moving into contango from its current backwardation condition.Let's assume that I'm using an n-factor intere...

<t>Are you asking for an extrapolation method to get implied vols at extreme strikes? Why? At these strikes I would think that the market dynamics of supply and demand play a far more important role than any number generated by clever models/extrapolation methods. Having said that, take a look at th...