1- TEST OF POSSIBLE IMPROVEMENTS
a- looser bhabha rejection
from:
reject if one electron > 65 GeV below 250
(above1550)
reject if acoplanarity< 30 and both
electr. > 40 GeV
to:
reject if one electron > 70 GeV below 250
(above1550) and ch.
mult. <13
reject if acoplanarity< 30 and both
electr. > 40 GeV and ch. mult. <13
b- more accurate selection of H-->
tau tau decays
from:
multiplicities of the tau "jets" < 4 and masses < 2 GeV
to: in addition, ask at least 1 odd multiplicity
c- alternative to b: give up to H-->
tau decay (a good reason for it: the second
discriminating
variable, i.e. btag)
Tested on 206.7C MC samples at the final cut for 2-d input
production
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
Our proposal: apply both a and c changes (b does not really
help)
This choice gets rid of the btag problem for H--> tau
tau decays
better efficiency and similar bkg compared to the
standard analysis
2- ARE WE MISSING EVENTS? CHECK WITH GUENNADI eeqq ANALYSIS
111223/3356 : in theta crack; E/p below 0.3 and dE/dx < 1.4
111851/1811: killed
by anti-Bhabha cut (fwd electr. > 65 GeV) -
recovered by new selection
112579/20089: one electron killed by isolation cut
112708/1581: sum of jets and electron masses below 150 GeV
115072/5758: killed
by anti-Bhabha cut (acoplanarity) -
recovered by new selection
115947/1106: one electron killed by isolation cut
116590/7648: one electron rejected because impact par. (z) above 1 cm
116629/16405:
one "electron" rejected because forward shower without
charged track
Conclusion: the events are not "lost" for mysterious reasons
but for features of
the analysis applied to MC as well as to real data. The
MC/data discrepancy may
be due to a fluctuation or to a very special kind of
incorrect MC description
affecting only a selected category of events.