Message-ID: <11459069.1075863444226.JavaMail.evans@thyme> Date: Wed, 8 Aug 2001 11:14:09 -0700 (PDT) From: tanya.tamarchenko@enron.com To: j.kaminski@enron.com Subject: Mime-Version: 1.0 Content-Type: text/plain; charset=us-ascii Content-Transfer-Encoding: 7bit X-From: Tamarchenko, Tanya X-To: Kaminski, Vince J X-cc: X-bcc: X-Folder: \VKAMINS (Non-Privileged)\Kaminski, Vince J\Inbox X-Origin: Kaminski-V X-FileName: VKAMINS (Non-Privileged).pst Vince, just wanted to summarize for you how I see the issues related to Credit Reserve and Var models: Credit Reserve: In order to address RAC's requests and questions related to this model Research needs to be able to compile and run the code. These procedures are not straight forward. Compilation and running can be done by using some scripts prepared by IT. Runs by different users can interfere with each other and with production runs. All this involves knowledge of Java, Corba, multi-threading, etc. and our IT's specific implementations, files locations.... We need IT support in doing this. I asked RAC (Debbie Brackett), IT (Winston, Steve Stock) for this. The reaction was not positive. I even guess they assume that I struggle for the "ownership" of the code. Our calls to IT guys are considered as interruptions in their work. _______________________________________________________________________________________ Var Model: "Convolution VAR" (Naveen's approach) is under way. At the end of this month (according to IT) it will more or less work. It is based on 60 factor model (which I suggested, but which was not implemented for our Monte Carlo model). There will be no way to perform proper testing of the "convolution VAR" even for normal distributions (in this case - Analytical VAR). There are no plans of such testing, as far as I know. What will happen is comparison with backtesting for our 5000 portfolios and then endless discussions which portfolios' VAR numbers are "better" or "worse" for the new method. ________________________________________________________________________________________ Thank you for your help, Tanya