Hi all,
reading the bad and looking for more useful plots has been an intresting
work. With Concezio I have investigated the nature of the discontinuities
in the relevant plots which we use for the optimization.
Now, with an improved fit to the Mx residual distribution, the
distributions are much more reliable:
see for Vub (fit -80MeV- ; mean distribu.; sig distr.)
http://www.slac.stanford.edu/~asarti/recoil/update/MNH69.eps
http://www.slac.stanford.edu/~asarti/recoil/update/Meandist_9.eps
http://www.slac.stanford.edu/~asarti/recoil/update/Sigdist_9.eps
& for Vcb (fit -80MeV- ; mean distribu.; sig distr.)
http://www.slac.stanford.edu/~asarti/recoil/update/vcb/MNH69.eps
http://www.slac.stanford.edu/~asarti/recoil/update/vcb/Meandist_9.eps
http://www.slac.stanford.edu/~asarti/recoil/update/vcb/Sigdist_9.eps

As you can see the plots are quite different from the last ones that I've
presented. This is expecially true for the last points of those plots
 (10 -> 14).
(Th first 10 points are still energy cuts, but going from 20 Mev to
110 MeV: The 80 meV cut is the 7th point placed in bin == 6
I've done this to remove the firts dummy cut E = 10)

This is because we have replaced the 12th point (cut in forward and bkw
regions) with a new cut:
12th point = Cut on energy (E>80MeV) && Cut on the number of neutral
reconstructed (n <= 2).
( See vub:
http://www.slac.stanford.edu/~asarti/recoil/update/MNH129.eps
& vcb:
http://www.slac.stanford.edu/~asarti/recoil/update/vcb/MNH129.eps
)

To understand the reason of this new cut you can see the follwing plots
showing the resolution on Mx (no kinematic fitting and no bkg subtraction)
for:
a) all candidates (white) && only chg particles (red)
http://www.slac.stanford.edu/~asarti/recoil/update/Mxreso_cont_ch.eps

b) all candidates (white) && only neu particles (red)
http://www.slac.stanford.edu/~asarti/recoil/update/Mxreso_cont_neu.eps

c) all candidates (white) && only events with 2 neutrals (red)
http://www.slac.stanford.edu/~asarti/recoil/update/Mxreso_compch2neu.eps

d) all candidates divided in variuos contribution:
http://www.slac.stanford.edu/~asarti/recoil/update/Mxreso_cont.eps

red = all events with <= 3 neutrals reco
green = all events with <= 2 neutrals reco
blue = all events with <= 1 neutrals reco
yellow = all events with == 0 neutrals reco (only chg part)

If you now see the optimization plots:
r : http://www.slac.stanford.edu/~asarti/recoil/update/MXEvents9.eps
Mx : http://www.slac.stanford.edu/~asarti/recoil/update/MEvents9.eps

Seems that this cut is powerful.
Of course the problem is that we are loosing a part of Vub events.
(Compare:
E == 80Mev cut
http://www.slac.stanford.edu/~asarti/recoil/update/MNH69.eps

E == 80Mev cut + Num of gamm <= 2
http://www.slac.stanford.edu/~asarti/recoil/update/MNH129.eps
)

If you mind why the cut on the gammas mother identity does not reflect now
the gain that we see with a simple number of gammas cut (<=2) you can see:
http://www.slac.stanford.edu/~asarti/recoil/update/Mxreso_comppiz.eps
that shows the contribution to the residual Mx distribution from events
with only gammas from pi0s.

The Fw and Bw regions are no longer considered because of cutting events
at low energy leads only to a worse statistics and does not gives a better
sigma or ratio (You can see the last point -bin = 14- in each of the above
plots...)

The possibility of applying this "number of gamma" cut has to be discussed
and, of course, needs further study mainly because of the gain in the not kinematic
constrained Mx residual distribution (saving a very good fraction of the
core events).

Any comment or question is welcomed.
Cheers,
Alessio

______________________________________________________
Alessio Sarti
 Universita' & I.N.F.N. Ferrara
 tel  +39-0532-781928  Ferrara

"Quod non fecerunt barbari, fecerunt Berlusconi"

"Che il bianco sia bianco, e che il nero sia nero
 che uno e uno fanno due e che la scienza dice il vero....
 DIPENDE !"