Frank, 

I don't think the issue is how the books are calced/officialized on those 
days.  Here why.  The same calc is used the day before and for the rest of 
the following month.  The physical calc is simply copied each day.  On Jan. 
31, extending the expert clause to include Feb is the only change made.  Once 
Feb rolles to intramonth the "From date" on the calc is changed from the Jan 
1 to Feb 1.  Since January has already rolled off this should have no effect 
on the positions.  Also, the positions for prompt month were correctly 
captured by Risktrac, which leads me to believe that problem is with the VAR 
calc not the post ids.  

Let me know if this doesn't answer your question. 
Robin  


   
	Enron North America Corp.
	
	From:  Frank Hayden @ ENRON                           02/07/2001 07:31 AM
	

To: Wei Hu/HOU/ECT@ECT
cc: Bharat Khanna/NA/Enron@Enron, Nilay Basu/HOU/ECT@ECT, Robin 
Rodrigue/HOU/ECT@ECT 
Subject: Re: INTRA-WEST-PH1  

Wei,
I appreciate your help. If I hear you right, it appears it may have to do 
with the way post id's and books are officialized.  Thanks for your help.

Robin,
For whatever reason, we didn't pick up VaR for Feb intra-month on Jan 31, and 
on Feb 1, a lot of intra month hit.  I've been investigating, with Bharat, 
and RAC initially thought it had something to do with the code and that GRMS 
didn't roll in physical until the first of the month (I guess).  Now it seems 
like it may have something to do with the way Post id's/books are 
officialized. (I guess).  Please give this a looksy and let me know.

Nilay,
Thanks for your help, any ideas on why would be helpful.

Let me/Bharat know if we should bring ERMS into the picture.

Thanks,
Frank



   
	
	
	From:  Wei Hu @ ECT                           02/06/2001 04:24 PM
	

To: Bharat Khanna/NA/Enron@Enron
cc: Frank Hayden/Corp/Enron@Enron, Nilay Basu/HOU/ECT@ECT 

Subject: INTRA-WEST-PH1

Bharat,

First I checked VaR number for this portfolio_id since 26-JAN-01.  Notice 
there was consecutive VaR jump from 26-JAN-01 to 1-FEB-01.

PORTFOLIO_ID    EFF_DT        DOWN95       UP95
--------------- --------- ---------- ----------
INTRA-WEST-PH1  26-JAN-01 -534098.14  468809.11
INTRA-WEST-PH1  29-JAN-01 -2472601.4 2351240.92
INTRA-WEST-PH1  30-JAN-01 -4860067.9 4509638.58
INTRA-WEST-PH1  31-JAN-01 -5771829.8 5412690.81
INTRA-WEST-PH1  01-FEB-01   -6776703 6805456.84
INTRA-WEST-PH1  02-FEB-01 -6328870.8  6454349.9
INTRA-WEST-PH1  05-FEB-01 -5535044.3 5546160.24

Then I checked sum of delta_position and gamma for book_id INTRA-WEST-PHY on 
those dates.  Since I notice that VaR had a big jump from 26-JAN-01 to 
29-JAN-01, so I also group by reference data on those two dates.

  1  select ref_dt, strip_flag, sum(delta_position), sum(gamma)
  2  from rms_open_position
  3  where eff_dt = '&1'
  4  and book_id = 'INTRA-WEST-PHY'
  5  and nvl(strip_flag, 0) >= 0
  6  group by ref_dt, strip_flag
  7* having (sum(delta_position) > 0 or sum(gamma) > 0)

Effective_dt 26-jan-01

REF_DT    STRIP_FLAG SUM(DELTA_POSITION) SUM(GAMMA)
--------- ---------- ------------------- ----------
30-JAN-01          3          152041.923          0
31-JAN-01          3          161992.376          0


Effective_dt 29-jan-01

REF_DT    STRIP_FLAG SUM(DELTA_POSITION) SUM(GAMMA)
--------- ---------- ------------------- ----------
31-JAN-01          3          153524.414          0
01-FEB-01          3          156324.294          0
02-FEB-01          3          156324.294          0
03-FEB-01          3          156324.294          0
04-FEB-01          3          156324.294          0
05-FEB-01          3          156324.294          0
06-FEB-01          3          156324.294          0
07-FEB-01          3          156324.294          0
08-FEB-01          3          156324.294          0
09-FEB-01          3          156324.294          0
10-FEB-01          3          156324.294          0

REF_DT    STRIP_FLAG SUM(DELTA_POSITION) SUM(GAMMA)
--------- ---------- ------------------- ----------
11-FEB-01          3          156324.294          0
12-FEB-01          3          156324.294          0
13-FEB-01          3          156324.294          0
14-FEB-01          3          156324.294          0
15-FEB-01          3          156324.294          0
16-FEB-01          3          156324.294          0
17-FEB-01          3          156324.294          0
18-FEB-01          3          156324.294          0
19-FEB-01          3          156324.294          0
20-FEB-01          3          156324.294          0
21-FEB-01          3          156324.294          0

REF_DT    STRIP_FLAG SUM(DELTA_POSITION) SUM(GAMMA)
--------- ---------- ------------------- ----------
22-FEB-01          3          156324.294          0
23-FEB-01          3          156324.294          0
24-FEB-01          3          156324.294          0
25-FEB-01          3          156324.294          0
26-FEB-01          3          156324.294          0
27-FEB-01          3          156324.294          0
28-FEB-01          3          156324.294          0


The following is the list of sum of delta position on those dates without 
group by reference_dt:

EFFECTIVE_DT STRIP_FLAG SUM(DELTA_POSITION) SUM(GAMMA)
------------ ---------- ------------------- ----------
   26-JAN-01          3          314034.299          0 (Reference date 
30-JAN-01 to 31-JAN-01)
   29-JAN-01          3          4530604.63          0 (Reference date 
31-JAN-01 to 28-FEB-01)
   30-JAN-01          3          8901557.37          0 (Reference date 
01-FEB-01 to 28-FEB-01)
   31-JAN-01          3          9539555.15          0 (Reference date 
02-FEB-01 to 28-FEB-01)
   01-FEB-01          3          8782018.97          0 (Reference date 
01-FEB-01 to 28-FEB-01)
   02-FEB-01          3          7889069.32          0 (Reference date 
06-FEB-01 to 28-FEB-01)
   05-FEB-01          3          7597454.74          0 (Reference date 
07-FEB-01 to 28-FEB-01)

It looks to me the trend on VaR change is very similar to the trend on delta 
position change.  If you want to know why delta position went higher from 
26-JAN-01 to 01-FEB-01, we need to get ERMS people involved.  RisktRAC 
basically just pulls position from ERMS.

Thanks,

Wei