Message-ID: <19943571.1075858477910.JavaMail.evans@thyme>
Date: Fri, 18 May 2001 14:32:32 -0700 (PDT)
From: kaminski@enron.com
To: vkaminski@aol.com
Subject: FW: Power VAR- comments on Winton's document
Mime-Version: 1.0
Content-Type: text/plain; charset=us-ascii
Content-Transfer-Encoding: 7bit
X-From: Kaminski, Vince J
X-To: 'vkaminski@aol.com'
X-cc: 
X-bcc: 
X-Folder: \Vince_Kaminski_Jun2001_10\Sent Items
X-Origin: Kaminski-V
X-FileName: vkamins.pst



 -----Original Message-----
From: 	Tamarchenko, Tanya  
Sent:	Thursday, May 17, 2001 4:01 PM
To:	Kaminski, Vince J
Subject:	Power VAR- comments on Winton's document

Vince,

1) I agree with paragraphs 1 and 2 in Winton's document (re-running jumps and clustering and correlations of jumps).

2) Regarding 3 (fixing instability in volatilities) I would suggest implementing volatilities smoothing (see short document attached)
 
3) Also: regarding correlations. The analysis I did was showing that correlations on fixed contracts prices are much more
consistent, have less noise versus prompt, promt+1, etc prices correlations (remember those pictures?). I was suggesting
to switch to those. IT implemented this in some version a while ago, but since they don't use version control environment properly
I am not sure they still have it. This will allow us to capture also the correlations across gas and power (I experimented with NG-R6 correlations) at least for 2 prompt years. 

4) We need to ask IT to increase the number of columns in the database table which holds factors from 7 to 100. Then we (Research)
can experiment with joint factor analysis to capture term-structure of correlations across curves.

Tanya