Hi Daniele,

> these values are not making sense. Eps_Cut ~ 1 means that you have only
> events < 1.5GeV. I presume that, for instance, in the hybrid case the fit
> is taking into account only the pure resonant part. That can explain why
> you have basically all the events < 1.5GeV.
>
> Please, check again the fit code. The way it is mixing resonant and
> non-resonant can explain the problem.

I saw (at last part of) the problem now. Thanks a lot for helping. By
default, the fit code takes only the resonant events from the hybrid MC.
This is not what I should do here. For now I just commented out the part
of the code doing the reweighting of the nonresonant events, such that
when I take the hybrid file only, the full hybrid file is taken (res and
nonres events). Also in both cases (only hybrid file and only nonres file)
no 3d reweighting is applied to the nonres events (no fermi reweighting
anyway or other vub specific reweighting), such that the nonres and
hybrid file are taken 'as they are'.

Virginia, should anything else be changed in the code apart from
commenting out what comes behind
      "//Star vub non-resonantspecific part"?
I did not see another place yet that needed changes for this.

I will give the numbers which I found here:

*** default
 BRBR                0.0219666  0.00234676  0.000873771
Eps_Cut = 0.775344 +- 0.00854322
Eps_tot = 0.248835 +- 0.00501357

*** nonres
 BRBR                0.0220791  0.00236418  0.00104411
Eps_Cut = 0.777663 +- 0.0137036
Eps_tot = 0.247003 +- 0.00801005

*** hybrid
 BRBR                0.0196832  0.00210833  0.000834273
Eps_Cut = 0.803137 +- 0.0103391
Eps_tot = 0.276976 +- 0.00683329

This does look better than yesterday, though not like 3% for the hybrid.

Here is the plot for the signal mc:
http://www.slac.stanford.edu/~kerstin/vubchop_onevubcomp.eps

Kerstin