To perform an analysis in DaVinci, several types of samples can be used, which represent tighter or looser restrictions to the selection of the particles.
The following notes are just a copy of the comments in the StandardDC06Options.opts file. The may not be valid for other Monte Carlo productions.
These sets of particles are useful for loose stripping and PID-studies. NoPIDsParticleMaker is used for creation:
Note that the NoPIDsParticlreMaker uses VTT, Long and Downstream tracks by default. See NoPIDsParticleMaker in the DaVinci Doxygen for details.
CombinedParticlMaker (with “ExclusiveSelection=false”)(comment on this) is used for charged particles with the definitions:
Containers:
See CombinedParticleMaker in the DaVinci Doxygen for details.
CombinedParticlMaker (with “ExclusiveSelection=false”) is used for charged particles with the definitions:
Containers:
(Cnv)PhotonParticlemaker is used.
See PhotonParticleMaker or CnvPhotonParticleMaker in the DaVinci Doxygen for details.
There are a set of StandardDC06 basic particle, StandardDC06 Pi0s and a set of StandardDC06 intermediate states: D0, Ds, D*, J/psi, K*, Phi. The relevant algorithms are fired on demand when one wants to read their outputlocation. See the relevant options for a complete list.
The cuts coded in this version are 'just a little tighter' than the ones applied in the exclusive HLT (with the big exception that PID is used here but not in the HLT). So hopefully events selected using these particles should also pass HLT. This has not been tested.
The actual definition of the cuts should be defined by the physics WG for the DC06 stripping. The present first import should be taken as an example more definitions of 'Tight' 'Looser' or 'NoPID' intermediate states could be provided if needed.
m_exclusive: Job option for exclusive selection of particle types.
PIDs are checked in some order and all tracks that are compatible with being a muon are made a muon, then an electron, proton, kaon and pion. What you get will depends on this order, which is already a good enough reason not to want that. If you do B→pipi and one pi decays to a muon, it will never be a pi candidate. Even worse, if you or someone relaxes the mu ID cuts between two jobs or two DaVinci versions, then suddenly what used to be a nice pion in one version of DV might now becomes a muon candidate. Your efficiency decreases although nothing affecting pion or kaon ID changed. That makes it very hard to estimate signal efficiencies as all efficiencies depend on everything. If you want to cut hard on PID, use PID cuts. If you don't want to use the same track several times (which is a valid point for tagging), then use the Overlap Tool and re-weight the track accordingly if needed.
The corresponding information for MC09 samples can be found in: $COMMONPARTICLESROOT/python/ in the DaVinci installation