1. Introduction
All results and input files are available at 


In the following (in no particular order), the updates for the winter
conferences 2003 with respect to Summer 2002 are discussed.

Note that all information is here  for discussion and not final. It is
provided "as is",  to have a starting point.  All  of your feedback is

All  information here is  preliminary and  confidential.  Some  of the
results have not yet been  approved by their collaborations, so please
do  not spread  the  numbers  here.  This  is  only for  HFAG-internal
discussion (and eventual approval).  Please keep this private.

2. Update of common.param
common.param contains the external input parameters, like lifetimes,
secondary branching fractions, etc. 

Modified to PDG 2002 values:
   dkp        3.80          +0.09      -0.09     ! D0->K+pi-   !mod
   dkpp       9.10          +0.6       -0.6      ! D+->K+pi-pi+ !mod
   dkpzp      13.1          +0.9       -0.9      ! D0->K+ pi- pi0 !mod
   dkppp      7.46          +0.31      -0.31     ! D0->K+ pi- pi+ pi- !mod
   dkzpp      2.96          +0.18      -0.18     ! D0->K0 pi+pi- !mod
   dnlx       6.87          +0.28     -0.28      ! D0->Xlv !mod

Modified to Olivier's latest values (2003), taken from 
   tb0        1.539         +0.014     -0.014    !x B0 lifetime !mod
   tbplus     1.656         +0.014     -0.014    ! B+ lifetime !mod

Added a parameter for f+/f0, the production ratio of B0/B+ at the
   fbpb0      1.072         +0.058     -0.058    ! Y(4S)->B+B-/Y(4S)->B0B0bar !mod
This is PDG 2002 with the old lifetime ratio. 

Using PDG2002 lifetime ratio (1.083), I get [fbpb0*(oldratio/newratio)]
   fbpb0      1.055         +0.055     -0.055    ! Y(4S)->B+B-/Y(4S)->B0B0bar !mod
Using 2003 lifetime ratio (1.073), I get [fbpb0*(oldratio/newratio)]
   fbpb0      1.065         +0.055     -0.055    ! Y(4S)->B+B-/Y(4S)->B0B0bar !mod

NB: o dkzpp changed from 2.70 -> 2.96. This is the most significant
---   change. 

3. f+/f0

The production ratio  of B+ vs. B0 mesons on  the Upsilon(4S) has been
traditionally set  to 1. This quantity  is now being  measured, and we
can  include   the  central  value  and   the  error  on   it  in  the
averages. This will affect B0 -> D*lnu and inclusive s.l. width on the

In the PDG, f+/f0=:r is given. The relevant parameter for the analyses
is f0*N(Upsilon(4S)). f0 = 1/(1+r) with the constraint that f0 + f+=1. 
Two setups were tested: 

 o Use f+/f0 with the error from PDG and use BABAR's systematic
   error for both BABAR and Belle. 

 o  Use the  calculated f0  (with errors  propagated) and  use BABAR's
   systematic error for both BABAR and Belle.

The  scaled  results  of  both  experiments are  compatible  to  three
significant  digits. The usage  of f+/f0  is probably  more fool-proof
since no  translation has to  be done. The  only thing to  remember is
that the change go into different directions (because f+/f0 != f0).

4. Updates of |Vcb|*F(1) and |Vcb|*G(1)

No new measurement here. Some of the results are rescaled slightly due
to the different input parameters, the average do not change.

The difference in f+/f0 between Belle and CLEO not (yet) corrected.

5. Update of BF(B0 -> D*+ l nu)
There is a new preliminary BABAR measurement.  

For  reference,  last  summer's  value  was  BF  =  (5.39  +/-  0.21)%
(CL=0.053).  Without  the new BABAR measurement, the  new common input
parameters yield BF = (5.41  +/- 0.21)% (CL=0.066). The largest change
is for Aleph, from 5.91% to 5.73%. This is due to the change in the BF
of D0->K0S pi+pi-.  The changes of the other measurements are smaller.

Including the new BABAR measurement we get the following.

a) With no correction for f+/f0:
   BF = (5.17 +/- 0.03 +/- 0.18)% (CL=0.0289).

b) With correction for f+/f0:
   BF = (5.27 +/- 0.03 +/- 0.19)% (CL=0.0638).

6. Inclusive semileptonic branching fraction

The following  changes have  been made with  respect to  last summer's

o The measurement of CLEO (PR  D45, 2212 (1990)) still included in the
  PDG 2002 average is not used since  it is based on a fit with prompt
  and  cascade shapes to  the lepton  spectrum and  does not  relie on
  event tagging as the other analyses.

o Inclusion of the two preliminary BABAR and Belle results with event
  tagging based on a fully reconstructed B decay. 

7. Average of inclusive charmless semileptonic branching fraction

This needs to be discussed. Four measurements on the Upsilon(4S) are

 CLEO endpoint   Phys.Rev.Lett.88:231803,2002
 BABAR endpoint  hep-ex/0207081
 CLEO mX-Q2      hep-ex/0207064
 BABAR mX        preliminary, to be approved. NOTE: central value "blinded"

There are two choices of results to combine: 

  o |Vub|  is given for  all analyses, albeit with  different formulae
    for BF -> |Vub| and  different errors. Would need consistency, but
    that can be done.

  o Inclusive charmless  branching fraction. This is not  given in all
    papers (absent in CLEO's  mX-Q2 analysis). Can be re-computed from
    |Vub|. Sometimes, not all errors  are given for the full branching
    fraction but  only for  |Vub|. In this  case, the  relative errors
    from |Vub| (x2) are applied to the branching fractions.

Some remarks on this first attempt:

 o b->c modeling is done with the same MC simulation (ISGW2, Goity
   Roberts, and FF parametrization) in CLEO and BABAR.

 o b->u modeling is done with the MC, again (including deFazio/
   Neubert for the non-resonant production).
 o f_u in the current BABAR endpoint analysis is fully correlated with
   CLEO's f_u.

 o The error of the applicability of b->s gamma to the endpoint
   analysis is common to both endpoint analyses.

 o  The  theoretical  error  in  the mX  (-Q2)  analyses  is  somewhat
   correlated  with  f_u,  as  OPE  is  used in  both  cases  and  the
   extrapolation  factor is  dependent  on the  same input  parameters
   (Lambda_bar,lambda_1 or mb,a). The  theoretical error and the error
   due to fu are assumed to be fully correlated.

 o The  central values used in  the non-resonant MC are  not stated in
   CLEO's hep-ex/0207064,  and they do not mention  the variation they
   applied to estimate the error.

 o In hep-ex/0207064, CLEO warns about statistical correlations
   between their two analyses. This has not yet been quantified. 
   No central value for the full BF is given. It would be useful to
   get this from CLEO directly. 

8. To do and remaining issues
The following will be updated in the next day or two: 

o s.l. width and |Vcb| with f+/f0