Problem reported in September Delphi week: loss of a candidate because of missed track/shower association.
Recovery procedure: force association of any track above 4 GeV/c to any shower (not belonging to other tracks) within a given angle (ANG_ASS) from the track.
Tested on several MC samples at 206.5 GeV:
Recovery tested on signal, q qbar, CC 4f and NC 4f for
ANG_ASS = 0.50, 10,
20, 30
and compared with the standard (0.0).
Results are presented in plots
.
The signal and signal-like background increase significantly for ANG_ASS until 10, then flatten. On the other hand q qbar and CC 4f have a tendency to explode after 10, so this value is used in the following analysis.
A data/MC comparison at preselection level is presented in figure . No major problem is observed.
A numerical comparison at the different analysis steps
is shown in the table (the contribution from Bhabha is 0 at preselection):
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
The results (and the problems) are generally the same as in the December 2000 analysis, except an increase of the qqbar background. On the other hand this is based on 12 MC events: