Print

---

Hi,

 I did changes in the fit. The main points are (the ones that
actually affect the final number):

 1) the efficiency for the signal is now calculated on the generic MC and
the value is 13% lower than the one on the cocktail.

 2) the ratio  factor eps_u^l/eps_sl^l is now calculated directly in the
fit and not fixed (as before). This is more coherent when you perform
stability check and breakdowns. The value now is ~3% up then before =>
shift up in the reult of ~3%.

 For the remaining important (but not crucial) changes I just append to
the mail the message of my commit.

 Tomorrow I will redo all the fits already in the BAD and I will post
them.

 Cheers,

 Daniele

> Distribution: CVSNEWS
>
> I've just committed some changes in my source files. The log message
was:
>
> Name: daniele  **********************************
> Update of /afs/slac.stanford.edu/g/babar/repo/IslBrecoilUser
> In directory
shire02:/afs/slac.stanford.edu/u/ec/daniele/devel/Ibutestnew/Is$
>
> Modified Files:
>       b0cuts.dat envfit.csh fitNtp.cc inputfiles.dat
>       recoilAnalysis.cc settingfit.dat
> Log Message:
>
>
> Many changes and lot of work
>
>                                                      Effect on the
measure
>
> fitNtp:
>
>  1) insert missing reweighting (B and D decays) in the
>     "other" component                                  NEGLIGIBLE EFFECT
>
>  2) final error does not contain MC statistics anymore   ERR DOWN 5%
>
>  3) in *results*.dat all interesting numbers now,
>     including the MC errors. THey are very useful to check
>     the fit results and to calculate error due to MC stat NO EFFECT
>
>  4) beautiful plots for the result now. Two bkg components
>     shown instead of signal and other comp                NO EFFECT
>
>  5) Vub cocktail + Vub generic for the signal shape
>     and Vub generic only for the signal. Since eff for
>     Vub generic only is 13% lower the final number will
>     be ~11% up                                            RESULT 11% UP
>
>  6) factor eps_u^l/eps_sl^l now calculated directly in the
>     fit (a fixed number before). This is more coherent when
>     you do cut variation and breakdowns. On the new
>     chains the value is ~3% up => result will be ~3% up   RESULT ~3% UP
>
>  Daniele