Hi David, Vera, and Marcello,

thanks a lot for your inputs.

In the meantime, we have implemented the form factor weights for B->Dlnu, 
B->D**lnu and the B->Xulnu channels. None of these changes our results in 
any significant way (as expected).

I added the results for the newest fit with taking the weight for the D** 
component from the fit to the depleted spectrum to the bottom of here:
http://costard.lbl.gov/~kerstin/vubunf/r22/spectrum/spectrum.html

To answer Vera's comment on using the B->D*lnu BF from the Breco sample: 
We do see a significantly better fit for the depleted spectrum if we use 
the higher B->D*lnu BF that we used for the analysis on release 18. The 
numbers and plots are on the same webpage (with the heading "B->Xclnu BF 
Reweighting According to What We Used on 18"). So I think the effect of 
this is basically achieved with the extra weight for the D**. It seems 
more-or-less a matter of taste which of the options we choose -- get a 
scaling for the D** or find a justification for using a higher D* BF.

Thanks again,
Kerstin

On Thu, 17 Dec 2009, Marcello Rotondo wrote:

>
>
> Hi all,
>
> I want just add my thought about the scaling we need to apply to the D* or 
> D** component to get a good chi2 in the final fit.
> The fact that fitting the depleted sample with the D* free, we need to 
> rescale the D* up by a large amount, does not means we are measuring a larger 
> B->D*.
>
> As shown at the Coll. Meeting, fitting the Depleted sample with the D** free, 
> you get also un "unreasonble result" because the fit prefer to reduce the D** 
> fraction by ~30% when you know there are missing components. But also in this 
> case we are going in the same direction: the D* have to increased. (I have to 
> say here that I agree with kersint and I would prefer to fit the D**).
>
> But I want to remind you that in the fit on the Enriched  sample (to extract 
> Vub) we have the total Vcb component free in the fit. The fit on the Depleted 
> say that the shape need to be adjusted, and the relative fraction of the 
> D*/D** have to be increased to get a good chi2. Otherwise the chi2 are bad. 
> That all we can say without other inputs.
>
> Cheers,
> Marcello
>
>
>
>
> On Fri, 11 Dec 2009, Luth, Vera G. wrote:
>
>> Hello Kerstin,
>> 
>> sorry for the delay in response.
>> I am somewhat hesitant to offer advice, since I am not
>> directly involved.
>> 
>> Here is my 5c of contribution to the discussion:
>> 
>> Clearly your rescaling of the D* rate is unreasonable.
>> I would fix the Dln, since this BF appears to be solid
>> and should not be rescaled.
>> The idea of rescaling the D* is based on the notion that it is the
>> largest contribution and its BF is poorly known.  So, if you scale D* up
>> do you downscale down scale D**??
>> Another approach is to fix the D*ln to the BF measured with the BABAR
>> Breco samples. But then how do you fill the rest of the BF?
>> DO we just leave a gap?  the appears to be difficult to accept,
>> but may give better results!
>> 
>> Without definite measurements we really are in trouble!
>> 
>> I am afraid, these are very useful comments.  Nothing you do not know!
>> 
>> Ciao
>> Vera
>> 
>> 
>> 
>> ============================================
>> Vera Luth
>> SLAC  - Stanford University
>> 2575 Sandhill Road, Menlo Park, CA 94025
>> Phone:  650 926 2702   FAX  650 926 2657
>> ============================================
>> 
>> -----Original Message-----
>> From: Kerstin Tackmann [mailto:[log in to unmask]]
>> Sent: Friday, December 11, 2009 8:51 AM
>> To: vub-recoil; Lopes Pegna, David; Luth, Vera G.
>> Subject: Re: Analysis status
>> 
>> 
>> Dear all,
>> 
>> should we interpret the silence as that there are no objections to what
>> we
>> are saying/proposing?
>> 
>> Best,
>> Kerstin
>> 
>> On Tue, 8 Dec 2009, Kerstin Tackmann wrote:
>> 
>>> 
>>> Dear vub-recoilers, conveners, Vera,
>>> 
>>> we have continued to look into the things we showed at the CM last
>> month and
>>> are getting to a point where we would appreciate feedback on several
>> points.
>>> It would be nice for us to have these clarified before we move on to
>> produce
>>> a measured spectrum with complete systematics, which we will need to
>> then
>>> tune the unfolding. I am hoping that settling on a strategy now that
>> is
>>> supported by the AWG might spare us duplicating work.
>>> 
>>> (1)
>>> Is the fitting of a scaling factor for the D** component from the
>> depleted
>>> sample a strategy we are happy enough with? I did try rescaling the D*
>>> (see the Nov 10 entry here:
>>> http://costard.lbl.gov/~kerstin/vubunf/r22/spectrum/spectrum.html),
>> but given
>>> that supposedly we know the D* better (forgetting about the spread in
>> the
>>> different results...) I think I have some preference for
>>> rescaling the D**, for which our knowledge is more limited.
>>> 
>>> (2)
>>> We evaluated another set of systematics to see what the discrepancy
>> between
>>> the fit results for the B->Xclnu scaling factor is that we fit on the
>>> signal-enriched and -depleted samples. We go down from 5.5 sigma (stat
>> only)
>>> to 2.3 sigma including a good number of systematics. This page
>>> http://costard.lbl.gov/~kerstin/vubunf/r22/spectrum/spectrum.html
>>> has results from a lot of the studies we performed. I am also
>> attaching a
>>> slide that shows the breakdown of the systematics on the scaling
>> factor
>>> difference. If we cannot go ahead with this at this stage, I would
>> appreciate
>>> suggestions for what other studies we should be providing.
>>> 
>>> (3)
>>> We are currently only using FF reweighting for B->D*lnu given that mX
>> should
>>> not be very sensitive to FFs. So the only sensitivity we have comes
>> from the
>>> El cut at 1GeV. In the r18 round of the analysis the uncertainties
>> from the
>>> B->D*lnu FF were so low that we did not even quote them. However, we'd
>> rather
>>> include additional reweightings now than later if there are strong
>> opinions
>>> that reweightings for other channels are necessary.
>>> 
>>> Thanks,
>>> Kerstin
>> 
>
>
> +------------------------------------------------------+
>   Marcello Rotondo
>   INFN Sezione di Padova, Dipartimento di Fisica
>   Via Marzolo, 8
>   35131 Padova - Italy
>   Tel  : +390498277047
>   Fax  : +390498277102
>   Email: [log in to unmask], [log in to unmask]
> +------------------------------------------------------+
>
>