Implementation problem with Andersen & Andreasen 's "volatile volatilities" in Risk 2002

JONAH · February 9th, 2006, 4:59 pm

Hi folks,I wonder anyone has the same difficulty when trying to implement the model from "volatile volatilities"(Risk 2002)? this model seems powerful, and has been used in Piterbarg's ground breaking paper "time to smile".The thing is: my numerical Fourier integral (follow the fomula presented in December Risk 2002, page 165) doesn't return good results(maybe I am wrong), but when I set parameters to let the model reduce to the simplest ATM "deterministic volatility" case (that is, skew parameter =1, vol of variance = 0, mean reverting rate = 0, variance at time 0 = 1)the Fourier integral should return 0 to get the Black price, but it doesn't....any comments are more than welcome.Cheers,

piterbarg · February 13th, 2006, 7:14 am

there is a typo in the original paper. Vladimir

JONAH · February 13th, 2006, 8:44 am

Thank you very much!!Jonah

piterbarg · February 13th, 2006, 9:03 am

finally managed to upload the corrected version, see attached-V

jaccker · May 12th, 2006, 2:46 pm

In this paper (pit's corrected version), equation (5), in the integral, it is (exp(-iw+0.5))*ln(&(0)/k) or exp((-iw+0.5)*(ln(&(0)/k))). I am confusing about it. Anybody could help me out?

GoGoFa · May 15th, 2006, 6:50 am

Your last choice is the correct one.

Atos · February 7th, 2008, 3:29 pm

I am trying to reproduce the formula in Preposition 1 from Andersen and Brotherton-Ratcliffe (J. Comp. Finance 2005), which I believe should be similar to those from Volatile Volatility. There seems to be the same typo. The characteristic function is not consistent with the definition in Lemma 2, and the argument of the function should contain -1/2 instead of +1/2. Did anybody reproduce this formula?Thanks.