Hoi,

we have a phone conference reservation for tomorrow for 1.5 hours:

   A MeetingPlace conference has been scheduled.
   Call: 510-665-5437
   When: September 25, 2003, 09:00 AM America/Los_Angeles
   Meeting ID: 2146

I have moving people coming  into my appartment at 10:00am (nominally,
maybe earlier, maybe later).

I append a draft of what we might use as a basis for discussion. 

Cheers,
--U.



   Hi Ed,

   thank you for your mail.   We appreciate your interest in our analysis
   and take your concerns very seriously.  Please find our answers below.


   > Also, you do not include an estimate of error for the assumption
   > that the two sets of parameters are IDENTICAL, rather than related
   > somehow in a way that is not understood.  I understand that not all
   > theorists feel as strongly as Neubert that your assumption that they
   > are identical is unreasonable, but don't they all feel that SOME
   > error must be assigned to this assumption?

   We  do include an  error for  the assumption  that the  parameters are
   identical. We  made a decision  (based on discussions  with theorists)
   not to use 90MeV as error on mb or LambdaBar (the agreement/outcome of
   the CKM  workshop 2002), but rather  use the full  error including the
   dominating  theoretical uncertainties (mostly  from higher  orders) to
   account for this.  It is  obviously a point of discussion whether this
   is an  appropriate or good  estimate of the  error and whether  or not
   this also covers the central value and error on lone or "a". 

   It is difficult  to assign meaningful errors to  unknown effects.  For
   example, "everybody"  agrees that "weak  annihilation" has potentially
   large effects in certain restricted regions of phase space, but so far
   nobody has determined an error  for this (e.g., in endpoint analyses).
   Quark-hadron duality  is in the  same category.  We are  therefore not
   setting a precedence by not  assigning an arbitrary error to something
   unknown.


   >     In our phone meeting of 3 July, the question was raised as to whether the
   > range of exponential shape function parameters (/\bar, lambda_1) gave
   > <E_gamma> values and errors consistent with CLEO's published values.  In Email
   > on July 4 I gave you the numbers showing that it did.  A correlary of this is
   > that if one calculates <E_gamma> using the values of (/\bar, lambda_1) that YOU
   > have been using, one will get an <E_gamma> value, with error, not in good
   > agreement with CLEO's measurement and error.  We have carried out this
   > exercise, and find:
   >                                             
   > Measured                                    2.346 +/- 0.032 GeV
   > calculated, using correct /\bar, lambda_1   2.353 +/- 0.031 GeV
   > calculated, using your /\bar, lambda_1      2.377 +/- 0.043 GeV
   > 
   > Note the differnce in the central value, 31 MeV.  Your values of /\bar and
   > lambda_1 come from our <E_gamma>, with its error, 2.346 +/- 0.032.  Your
   > procedure should, MUST get that back, to an accuracy small compared to the
   > statistical error on the measurement (32 MeV), if using the HQET /\bar and
   > lambda_1 in the shape function is reasonable.  You don't, and the only possible
   > interpretation is that HQET /\bar, lambda_1 don't mean the same thing as
   > Light Cone /\bar, lambda_1.  This conclusion is TOTALLY INDEPENDENT of what any
   > theorist tells you, or us.  It's an empirical fact, the output must match the
   > input, if what you're doing is right, and it doesn't.

   The parameters (LambdaBar,  lone) that we use are  derived from CLEO's
   publications of  both photon spectrum AND hadronic  mass moments (else
   there would be  no lone).  Your argument is  therefore not valid.  The
   fact  that <E_g>  in the  first and  third row  agree even  within the
   *statistical* error is actually  supporting our position that there is
   no evidence that the HQET  and SF parameters are different (within the
   precision currently achieved).

   To avoid  a misunderstanding:  Your calculation is  based only  on the
   two-parameter  shape function,  is that  right? Assuming  that  is the
   case, the agreement  between the first and second  row seems a (lucky)
   coincidence, as  subleading corrections will have an  effect. Is there
   any reason to assume that the  photon spectrum is controlled only by a
   two-parameter  shape  function?    Furthermore,  the  HQET  parameters
   (LambdaBar,  lone) are  certainly prone  to higher  order corrections,
   which are not  expected to be covered by  the statistical error alone.
   From all this we would  conclude that the results should be consistent
   only within the statistical  + systematic + theoretical error (whereas
   they are already consistent within the statistical error alone).


   >     If you choose not to use CLEO's determination of the shape function, but
   > instead use CLEO's determination of HQET parameters /\bar and lambda_1, as you
   > have done, then shouldn't you include SOME error for this approximation?  I
   > think the answer is a clear "Yes".  How big?  Hard to say.  Would it matter?
   > YES!  This is your dominant error.  You're clearly better off having it
   > determined by an objective procedure than by a guess.  And you're clearly
   > better off having it determined by a guess than by setting it equal to zero!

   We disagree here.  The relationship between the HQET and SF parameters
   is about  as unknown as weak  annihilation and QHD, both  of which are
   commonly not assigned errors.


   >     I recall, and Dave Cassel recalls, that you mentioned that the range of
   > uncertainty in /\bar, lambda_1 that we obtain from our b -> s gamma measurement
   > gives a varaiation in M_X distribution that you find in poor agreement with your
   > measured M_X distribution.  If this is true, if you indeed can reduce the errors
   > on the shape function parameters, that is great!  It certainly deserves mention
   > in your paper, since we're talking about the dominant error.

   This  measurement will  be the  topic of  a future  paper. We  feel we
   allude to this in the concluding section.


   > Finally there is your third error, for which you quote +/-0.40, but
   > I say you should quote +/-0.60, 13.0%.  ("I say you should" means
   > that's what you would get if you used the light cone /\bar, lambda_1
   > from our b -> s gamma determination.)  That is to be compared with
   > our second (+/-0.44) and fourth (+/-0.24) error, => +/-0.50, 12.3%.
   > These errors, yours and ours, in addition to being comparable (13%
   > vs. 12.3%), are HIGHLY CORRELATED.  So, collecting errors, 8.4%
   > vs. 8.3% (uncorrelated), 5.6% (completely correlated), 13.0%
   > vs. 12.3% (highly correlated).  The correlated errors dominate.
   >     Conclusions from this part.
   >         1. Your statement of smaller systematic errors is incorrect,if you
   >             include theory errors in the systematic errors.

   You use  "theory errors" to varying  degrees.  Why do  you not include
   errors for  quark-hadron duality or  weak annihilation?  Both  of them
   are expected  to be  significantly larger in  the endpoint due  to the
   much  smaller  phase  space.   Should  the  error  due  to  subleading
   contributions to the shape function be included?


   >         2. Putting all the errors together, your analysis and our analysis have
   >              very nearly equal errors, if theory errors are treated in a
   >              consistent way.
   >         3. There is very substantial correlation between your errors and ours.

   Much  more than  just  the  error for  BF->|Vub|  is correlated.   The
   b->clnu modeling is also correlated, for instance.


   Cheers,
   BABAR