Winston,

Thanks a lot. 

I talked to Tanya about testing VaR based on 20-day historical fwd-fwd vols.
It would be great if we had this done by the close of business today.

We should also push ahead to get the ability to use the fwd-fwd vols input 
from
different sources (historical, trader's input, etc.)

Vince





Wenyao Jia
06/05/2000 08:36 AM
To: Grant Masson/HOU/ECT@ECT, Tanya Tamarchenko/HOU/ECT@ECT, Vince J 
Kaminski/HOU/ECT@ECT
cc: Naveen Andrews/Corp/Enron@ENRON, Stephen Stock/HOU/ECT@ECT 
Subject: VaR problem with STORAGE


Hi everyone,

By using simulated curves and positions, I manually calculated both days's 
VAR for AGG-STORAGE in Excel spreadsheet.   I got almost same results as the 
ones produced by the current production VaR model.

I did found two things:

1.  The positions in Storage-Prc changed although the total position did not 
change much:
 For July delivery:  delta changed from 2,682,338 to 5,695,589
 For Sep. delivery: delta changed from 4,629,638 to 1,682,515.

2.  The forward forward NG volatility has a big change on Dec-01:
 May 31 :   0.0122956
 May 30 : 0.278
      But the Original Vol curve don't change much.


I also changed code so that it will read NG forward forward vol curve from a 
file.  We can run some testing today.

Look  like Jin found some problem in the factor loading  program.  If we can 
produce new factor loadings we can also rerun VaR using the new factor 
loadings.

If there is anything else we can do, please let me know.

Thanks!

Winston Jia