Volatility Surface

[aligr]

Hello everyone,

I am a litle desperate about transforming given deltas to strikes at the following "Volatility Surface".

10DPUT ... 25DPUT ... ATM ... 25DCALL ... 10DCALL
1W 14.481 ... 14.387 ... 15.250 ... 17.387 ... 19.281
2W 14.831 ... 14.737 ... 15.600 ... 17.738 ... 19.631
... ... ... ... ... ... ... ... ... ...
... ... ... ... ... ... ... ... ... ...
2Y 15.638 ... 15.825 ... 17.300 ... 20.225 ... 22.675

I have read Mercuiro's paper. According to paper (as far as I understood) I am
supposed to get one strike for each delta. But, I am getting different strikes
for each maturiy. What I want is exactly the same matrix as I submitted above except
the first row, instead of deltas I want to show the strikes. Any help?

 

[aligr]

You might need to explain yourself a little better. If you explain why you need this it would help, as I suspect that will reveal a fairly basic lack of understanding regarding the information in front of you.

Start from the beginning. What instruments are you trading? Why do you need a vol surface?

Will try and help you along, but if I just type the answer you probably wouldn't understand it. Small steps to get there is better I think.

GJ

Thanks for reply,

I am trying to price F/X options but I have to use the implied vols.

 

[aligr]

Ok - lets try some more digging......

Why are you trying to price FX options? And what's wrong with using vol? That's what the whole market uses.

I am trying to implement VaR system for F/X option positions which uses historical simulation method. House not wiling to use the historical vol but implied vols. So each time I run the simulation I need to price to F/X Options.

 

[aligr]

and the reason I was asking about what instrument you are trading is to determine whether this is an exchange traded or otc portfolio of options you're talking about. edit - this is relevant when we talk about strike.

Is this for a bank? All seems a tad amateurish to me mate.

OTC options.

 

[jamescummins]

To slightly echo what Gamm said, if you are trying to build an in-house VaR model, what are you using as variables? Do you want to measure you risk associated with current implied volatilities? A robust VaR model will be tricky to build for any OTC intrument, and to be brutally honest, I would explain to your employer that you either really need them to be very very specific about what they want (I suspect they are not looking for a true risk model from you!), or to ask their quants to build it instead

I dont know the source, but is the matrix you first posted built from implied, or historical volatilities? If you want to build a model from the historical implied volatitlies, that could actually be interesting Never really thought about it, but I would be keen to know how the implied surface compares with the realised historical one. I would imagine over time they tend to the same shape. Also, what sort of historical period are you using? What sort of timeframe is the model for? Realtime?

I am a nerd, and problems like this interest me, but I dont really understand what you want this model to do...