Re: [Quantlib-users] Query on gamma

[Ashish B. <ash...@gm...>]

Hi Klaus,

Thanks for response. Where could I read about our get example of what you
suggested:

"one or two damping steps or use the TrBDF2 scheme."

By damping step do you mean by shocking the underlying price revaluating
the option?

Regards
Ashish

On Sat, Nov 19, 2022, 3:50 PM Klaus Spanderen <kl...@sp...> wrote:

> Hi Ashish,
>
>
>
> the gamma is in absolute terms, meaning just the second derivative of the
> npv w.r.t. the underlying. By default QuantLib is using Crank-Nicolson
> scheme, which is fast but might give instable gammas for some ATM options.
> To stablize the gamma you can either use one or two damping steps or use
> the TrBDF2 scheme.
>
>
>
> best regards
>
> Klaus
>
>
>
> On Donnerstag, 17. November 2022 13:31:29 CET Ashish Bansal wrote:
>
> Hi,
>
>
> I am calculating the option gamma using the following functions:
> engine = FdBlackScholesVanillaEngine
> Process = BlackScholesMertonProcess
> payoff = PlainVanillaPayoff with american option
> gamma function = option.gamma()
>
>
> For a few of the options, I am getting the gamma more than 1, up to 13. I
> want to know if the above function returns the gamma in percentage terms or
> decimal. Should I be dividing the gamma by 100 to get the correct value?
>
>
> Regards,
>
> Ashish
>
>
>
>