Using the event-by-event primary vertex finder PrimeVtx 
in the secondary vertex finder SecVtx increases
both positive and negative tag rates, as seen in the Summer 2003 
conference results (relative to the results shown in
the Winter 2003 conferences). 
The positive tag rate increased by around 40\% 
and the negative tag rate increased by around 93\% for
{\tt JET50}, while the $t\overline{t}$ efficiency for Monte Carlo
increased by 20\%.

While the increase in positive rates was expected,
there were some questions as to why the negative tag rates increased
by about 93\% relative to the Winter 2003 result. 
In order to answer this, we explore the usage of better beamline 
measurements and the introduction of PrimeVtx as the primary vertex
estimate. The beamline estimate used for Winter 2003 physics analyses 
overestimated the
beam widths by about 20\%, and this is a contributing factor
to the raise in negative tag rates. Also, the addition of {\rm ISL}
in {\tt TrackRefitModule} increased both the positive and negative
tag rates.

Tracking changes are responsible for an increase of
27\% in the negative tag rates, 
the more correct beamline measurement is responsible for an additional 20\% 
increase,
and the inclusion of PrimeVtx is responsible for an additional 46\% increase, 
for a total of 93\%.


Posted to /cdf/pub/cdf6666_primevtx_secvtx.ps
    http://www-cdf.fnal.gov/cdfnotes/cdf6666_primevtx_secvtx.ps

Posted by rappocc@cdfsga