[Quantlib-dev] VarSwap Implementation

[animesh s. <ani...@gm...>]

  I was trying out the Variance Swap (test-suite) implementation and 
came across some pricing issues.
Let me state this ->
1. The simplest way to price a variance swap is to simulate the stock 
using a random walk.
2. Calculate Log(Si/Si-1) at each step
3. Calculate total variance and annualize it
4. For several simulations repeat the above steps and calculate average 
variance
5. Square root of average variance is price of Variance Swap

Attaching a simple implementation based on above steps.

Next I proceeded with quantlib and used the method testMCVarianceSwap
t = Expiry time, I changed it to 1 instead of 0.24675

I was confused with parameters t1 and v1.
I assumed t1 can be either 1. Time from expiry, so setting t1 = 0 would 
have been good to go, but it resulted in error
                                           2. Time remaining before 
expiry so setting t1 = 0.99 would have been good to go, but again it 
resulted in error.
Anyway I tried it for t1 = 0.01 and t1 = 0.99 and it worked( Although I 
still dont understand why t1 can't be 0 or 1)

Same confusion with v1 and v. Couldn't find anything in documentation, 
so a bit confused.
For 20% vol the price came out to be very close to 20%.

 From my excel sheet the price is close to 31%.

Also I don't see any point in having a varstrike parameter, if we are 
calculating fair strike for a variance swap.
Can someone guide me in how it is implemented in Quantlib?






-- 
Regards,
Animesh Saxena

(http://quantanalysis.wordpress.com)
Ph: (+91)9920098221