Tanya,

Some stability tests were performed on the simulation.
(1) The Window size of the averaging in the simulation was changed from 10 to 
20 for NG-PRICE(the biggest book in gas) for effective date June-28 .  As you 
can see in the file 28NG-Price_Windowsize, the Component VaR numbers are very 
similar. 

(2)  To look at a calculational comparison, I looked at the STORAGE-PRC book 
(which has no gamma position) for effective date of 27th and (a) calculated 
FFVols, and (b)calculated the RiskMetrics VaR (the correlations are very high 
across the term structure) and compared to the Component Var code 
calculation, and again the two nubers are comparable, given the different 
modes of calculation.

Naveen