Hi Francesca,

> Hi Henning, an addition. If one does the global fit in the BLNP scheme and
> adds the results from the fit to the spectrum (within which scheme they
> are consistent), the errors on the parameters are going to be very small.

I'm not sure what you mean by global fit in BLNP scheme. There exist no
calculations for b->clv in this scheme and I'm pretty sure they won't for
a while. What Neubert is refering to is to use the calcualtions in
the kinetic scheme, relate the parameter definitions to those in the SF
scheme and then re-expand the expressions consistently in alpha_s and
1/mb. But this is pretty much the same as to fit in the kin scheme and
then translate the results too the SF scheme.

> That has not been yet tried. Note moreover, that
> the fit to the spectrum in BaBar gives results much better than from
> Belle.

But here we again encounter the problem of model dependence. The Belle
spectrum is much less sensitive to the influence of the K* peak as it is
measured in the Y(4S) and therefore this is smeared out and so I would
not expect that from a fit to the spectrum you get the same results. The
moments however which qre not so sensitive to the differential spectrum
agree pretty well. The fact that the fit to the spectrum from the Babar
semi-excl analysis agrees so well with the clv moments could just be a
coincidence.


Cheers,
Henning

> On Fri, 29 Jul 2005, Henning Flaecher wrote:
>
> >
> > Hello,
> >
> > If you use a fit to the spectrum you should try out several different
> > ansaetze for the shape function as a priori it is not clear what the
> > functional form of the shape function is.
> > For example, you can see that the results from the fit to the semi-excl
> > bsg spectrum change by ~1 sigma when using the Gaussian instead of the
> > Exponential ansatz. Of course it's a matter of discussion how to translate
> > this into the error on Vub.
> > In addition there are remaining uncertainties from e.g. the order of the
> > pert. calculation but this is more difficult to quantify.
> >
> > Concerning the fit to moments the theoretical uncertainties come from the
> > precision of the expression for the moments, i.e. higher order terms etc.
> > Here you integrate over the whole spectrum and so are less sensitive to
> > what happens in the resonance region.
> > For moments with high Ecut, the authors of the kinetic scheme calculations
> > add an additional theory error for the bias corrections.
> > However, in combination with the clv moments this leads to small
> > uncertainties on the SF parameters. The results for a fit to all bsg and
> > clv moments are summarised in hep-ph/0507253
> > I made a similar fit using only Babar input for the endpoint analysis.
> >
> > Cheers,
> > Henning
> >
> >
> >
> > Hi Wolfgang, Neubert supports the fit to the spectrum for both the old
> > calculation (which is commonly referred as 'Kagan&Neubert') and the
> > current BLNP calculation. It is not true that the fit to the moments is
> > more accurate, it is just different and interesting as well. For the
> > future Vub paper, quoting from an exchange of emails we had with Matthias,
> > he says: "However, I still think that
> > for the Vub extraction it may be better to follow the route using the
> > generator, since the B->Xu l nu spectra are predicted with the same theory
> > at NLO."
> > Where for 'generator' he refers to the notebook we got, and for
> > b2sgamma then this means to use the differential BF.
> > I am happy to discuss further this topic... the fit to the spectrum is
> > correct and supported by the authors and the parameters obtained should be
> > used for the extraction of Vub.
> > 		Francesca
> >
> >
> >
> > On Thu, 28 Jul 2005, Wolfgang Menges wrote:
> >
> > > On Fri, 29 Jul 2005, Concezio Bozzi wrote:
> > >
> > > > 1) We have been asked by the referee of our paper hep-ph/0504071 to
> > > > comment on the sentence:
> > > >
> > > > "However, there are concerns that the extraction of the shape function
> > > > parameters from the photon spectrum in B->Xs gamma is less reliable"
> > > > (p.18 before section 7)
> > > >
> > > > Could you explain to what you are referring to?
> > > >
> > >
> > > Actually, Vera and I have added this and, as written, it should hint to
> > > the fact that the extraction of the shape function parameters from the
> > > moments are more reliable than from the shape. To be more explicitly,
> > this
> > > should be seen in the context that we are using the shape function
> > > parameters extracted from the Belle b->sgamma spectrum [hep-ex/0407052]
> > > which uses the Kagan-Neubert prescription. To my understanding
> > > Kagan-Neubert is not as sophisticated as BLNP but I am not sure what the
> > > authors have said about the reliability of their calculations.
> > >
> > > We would have prefered shape function parameters extracted from moments
> > of
> > > the b->sgamma spectrum but this wasn't finished in time for LP05. For
> > the
> > > paper this will be different.
> > >
> > > Cheers,
> > >
> > > 	Wolfgang
> > >
> >
>