Object Selection:

Jets

More details to be found here: TopJetLiaisonR172Recommendations

• AntiKt 0.4 TopoCluster jets with: LCW+JES calibration. Use the AntiKt4LCTopoJets collection (D3PD name jet_AntiKt4LCTopoJets_bla e.g. jet_AntiKt4LCTopoJets_eta ).
• electron-jet duplicates
  • Remove one jet closest to an ACCEPTED electron (i.e. after final electron selection criteria - defined below) if it is within dR<0.2.
  • To compute the dR
    • Use electron track eta and phi
    • Use eta and phi of fully corrected jets using the ApplyJetCalibration tool.
  • For this duplicate removal step, consider all jets of positive energy jets in the collection (jets remaining after rejection of negative energy jets - no other cuts).
• the jet cleaning should be applied after all the event selection to both data and MC.

• • For the eta acceptance cut, use the fully calibrated jet eta (no origin correction is applied in 2012).

• • The MET uses a cut at 20 GeV for jets and 10-20 GeV jets for "softjets"
  • The muon isolation requirement uses a cut at 25 GeV (see here)
• When where is a "Bad" jet in the event we remove whole event. See event level cuts below for details.

Jet Vertex Fraction

We suggest to cut at: |JVF| >0.5 (wrt 0.75 for 2011, allows us to be consistent with other groups and b-tagging calibrations)

Further studies on the cut values, SF and possibly additional systematics are being carried out by the JetEtmiss group.

Outstanding problems

If electron is reconstructed inside a real jet, the energy is double counted. One solution is to remove electron clusters (or cells) and re-do jet reconstruction. This is a longer term task. Note that "simple" solutions, such as vetoing such miscreconstructed events, creates new problems and does not seem to be an improvement. See e.g. this thread. The current recommendation is to do nothing about the energy double counting. This is fine as long as the effect is reproduced by the simulation.

Jets for analysis

Jets used for the analysis are AntiKt4Topo jets. For analysis aiming for Top 2012, we stick with EM+JES calibrated jets. Everyone else should move already now to LC jets.

JES calibration

General information on re-applying the jet calibration is at JetCalibrationToolsForPhysicsAnalyses and JetStudies2012.

Note that for LC+JES jets (e.g. AntiKt4LCTopo), make sure that instead of the emscale variables you are using the constscale variables. This is because the scale at which the JES has been derived already assumes that the clusters have been calibrated with the LCW scheme.

Also note that currently, origin-corrected eta and phi should not be used.

Note that after the jet calibration is applied, the entire 4 vector of the jet can change. We propagate the change in pt to recalculate the MET, but keep the jet direction the same.

JES uncertainty

More information on how to use the uncertainty provider tool is available at JetUncertainties2012.

You should always use calibrated quantities after the ApplyOffsetEtaJES macro to evaluate the JES uncertainty, and rescale the energy components of the four-vector (E and pT, but the mass will change as well) using the uncertainty factor.

JER smearing and uncertainty

No smearing of the nominal MC jets is needed; smearing needs only to be done to obtain the systematic uncertainty on your final measurement.

Use ApplyJetResolutionSmearing-00-01-00 and JetResolution-02-00-00. Further instructions at JetEnergyResolutionProvider2012 and ApplyJetResolutionSmearing.

Jet Cleaning

Jet cleaning has to be applied to both data and Monte Carlo, and it is recommended to use the IsBadLooseMinus flag (=branch) in NTUP_TOP.

Events should be rejected if there is a bad jet with pt>20 GeV. For the sake of consistency among analyses we recommend to cut on the fully corrected jet pt after recalibration (i.e. the pt used in the analyses)

Further recommendations are available at HowToCleanJets2012.

JVF cut

Last recommended tag

• Pythia-derived scale factors
• Herwig-derived scale factors
• Averaged scale factors

NOTE : To obtain the last set of SF, you have to choose the year argument as 2012mc12a (and not 2012 anymore).

IMPORTANT! The SFs have been extracted after applying the additional mu re-weighting described in the InDetTrackingPerfo page. It is very important that you make sure you apply that to your analysis, too, before using the SF.