<t>Thanks for the reply seppar.Frans de Weert 2006 (section 7.4 p86) gives an alternative explanation. Summary: take the gamma/theta relationship theta=-1/2vol^2*S^2*gamma, insert expression for both gamma (N'(d1)/S*vol*sqrt(T)) and theta (S*vol*N'(d1)/2sqrt(T) [2nd term disappears when r=0 for dail...

How about flexing each of the vol inputs in your model and then calc the usual v(vol+dvol)-v(vol-dvol)/2dvol/100 (eg London(2005)p143)?

<r>Not sure if you still need those but have you looked at:<URL url="http://math.uchicago.edu/~sbossu/CorrelSwapPricing.pdfhttp://www.nuclearphynance.com/User%20Files/1522/BNP%20Paribas%20-%20Index%20Variance%20Arbitrage.pdfhttp://www.classiccmp.org/transputer/finengineer/%5BBNP%20Paribas%5D%20Volat...

<t>You're long a 1 year variance swap and the underlying moves 10% on one day and nothing for the remainder of the year. How much money do you loose? (Pretty much the entire vega notional? Realised var will be 10^2/255=0.392, assuming strike of say 20^2 [those were the days!] the long payoff is -399...

<t>I've heard this several times now: if you're long gamma and delta neutral (I think), how much does the underlying have to move on a daily basis so you don't loose any money (from time decay)? Supposedly the answer is the daily vol but I can't figure out why. Anybody have some insight please? Than...