Hello Bob,

[log in to unmask] wrote:

>I've been skimming bad1214 (I haven't read it carefully yet) and have a
>coupole quesitons:
>
>0) Why do we have a smaller SF parameter error than Belle, yet use SF
>input (mb) that has a larger error than what they use?
>
>  
>

Are you refering to the q2-mx analysis in the BLNP approach and HQE fits 
(8.2% error)? In this case I took what we showed at CKM2005, (8.2% to SF 
and 4.9% on theory) which was based on Matthias's Mathematica notebook. 
This might not be probably the most recent evaluation...any comments on 
this?

>1) Did you look at cuts other than 1.55 GeV for Mx?  If not, why not?
>
>  
>

Given time and manpower constraints we put more emphasis on the 2D 
fit... I agree, we have to try this as well.

>2) If you are going to quote a systematic due to "binning", please read
>http://babar-hn.slac.stanford.edu:5090/HyperNews/get/Statistics/217.html
>and follow-ups and explain why you think you're actually measuring
>anything of value by doing this.
>
>  
>

That 2.5% due to binning is the maximum variation wrt the default 
binning. I agree it's not the proper way to do it. Antonio run a number 
of tests by varying the binning, you will find results in

http://www.slac.stanford.edu/~petrella/tempo/system_bin.html

Given the statistical error and other systematic uncertainties, I think 
that effects due to binning are not large. In retrospective, I am afraid 
the 7% variation we saw in the Mx analysis last year is completely out 
of sense...

>3) Why do you not quote a partial BF for the MX analysis?
>
>  
>

We still need to thwart the code (we already have something), it would 
not be in time for LP...

>4) Figure 17 is fairly convincing...that there's a problem with the
>background estimation.
>
>  
>

Figure 17 is the 2-parameter fit where we fix the relative vcb and other 
components to the value determined on MC. Now we are fitting the Mx 
distribution with 3 parameters (N_vub, N_vcb, N_other), see fig. 18. It 
seems that events other than vub and vcb are less than expected on MC. 
We are trying to understand if the residual discrepancy beween 2.2 and 
2.5 GeV is due to higher D resonances.

>I'm eager to get these results into Francesco's talk; otherwise Belle will
>have a better Vub.... however, for publication I insist that Mx be treated
>with as much care as q2-Mx.  It would also be nice to have P+....
>
>  
>

I think the strategy here is to wrap up what we got up to now and get it 
out for LP. Publication will follow after EPS.
Concezio.

>Bob
>
>\\\\\\\\\\\\\\\\\//////////////////\\\\\\\\\\\\\\\\\//////////////////
>/ Robert V. Kowalewski            \/  Dept. of Physics and Astronomy \
>\ particle.phys.uvic.ca/~kowalews /\  University of Victoria         /
>/ Tel:   (250)721-7705            \/  P.O. Box 3055                  \
>\ Email: [log in to unmask]         /\  Victoria, BC V8W 3P6           /
>/////////////////\\\\\\\\\\\\\\\\\\/////////////////\\\\\\\\\\\\\\\\\\
>
>  
>