3.5.1

Update with improvements for stored generator-level weights.

• Allow to select which LHE and PS weights to save.
• By default, only store LHE weights for scale variations and PS weights for main scale variations. To store PDF weights, they should be specified in the configuration.
• Save mean values of PS weights before the event selection, similarly to what is done for LHE weights.

Renamed branch MeanAltWeights → MeanAltLheWeights in EventCounter. Apart from this, the format of produced files has not been changed.

3.5.0

First release to run on both 2016 and 2017 data.

This version uses CMSSW_10_2_X and is intended to run on the following MiniAOD data sets:

• 17Jul2018 data and RunIISummer16MiniAODv3 simulation for 2016.
• 31Mar2018 data and RunIIFall17MiniAODv2 simulation for 2017.

Running on data and simulation for 2018 is not supported yet.

Other changes besides the update in the release and data sets:

• Updated jet ID for 2017.
• Switched to Fall17_V2 ID for electrons in 2017 and photons in both years. Dropped custom electron ID without isolation requirements.
• Updated JER factors for 2017.
• Updated MET filters.
• Enabled saving of event weights for main parton shower scale variations. However, these weights are not available in 2016 samples and included only in selected 2017 samples.
• Enabled storing of event weights to account for L1T ECAL pre-firing in 2016 and 2017.

Format of produced files has been changed in a backward-compatible way.

3.4.0

First release to run on 2017 data and corresponding simulation.

• Updated to CMSSW_9_4_3.
• Electron and photon ID updated to version Fall17.
• Jet ID updated to preliminary 2017 version (presented here).
• Store electrons with p_T down to 15 GeV (changed from 20 GeV).
• In simulation store pileup profile before event selection to address problem reported here.

Format of produced files has not been changed.

3.3.4

Release to run on the legacy re-reconstruction of 2016 data, MiniAOD version 1. Note that this version is intended for POG studies, while physics analyses are supposed to run over MiniAOD version 2.

• Main configuration now supports flag isLegacy2016. When set to true, the configuration is adjusted for the legacy re-reconstruction. Switched off be default.

Format of produced files has not been changed.

3.3.3

Update to CMSSW_8_0_28 and a clean-up.

• Drop CSV b-tagger and c-taggers. Also drop jet pull angles.
• Allow a simple selection on jet b-tagging discriminators in the configuration. Documentation provided here.
• Add scripts to perform physics validation of releases.

This release should not be used for processing of the legacy re-reconstruction of 2016 data because it implements the MET corrections for the ECAL gain switch.

Format of produced files has not been changed.

3.3.2

Update with improvements for pileup.

• Store the number of “good” primary vertices instead of all of them (issue #98).
• Allow storing the largest p̂_T of admixed pileup interactions in the nominal bunch crossing. Disabled by default.

Format of produced files has been changed in a backward-compatible way.

3.3.1

Update that allows propagation of individual JEC variations into MET starting from PEC tuples.

• Jet pt threshold has been lowered from 20 to 15 GeV, which allows to recompute T1 MET corrections starting from information stored in PEC tuples.
• MET in data is now recomputed instead of reading slimmedMETsMuEGClean directly from MiniAOD. This fixes a rare problem in correcting for ECAL gain switch, which has impacted MiniAOD.
• In script crabSubmit.py it is now possible to choose what tasks should be submitted.

Format of produced files has been changed in a backward-compatible way.

3.3.0

Release to run on 03Feb2017 re-MiniAOD data and Summer16 simulation. It corrects several important bugs that affected 3.3.0-rc1.

Bug fixes:

• MET uncertainties in simulation have been corrected (issue #95).
• Filters for events with duplicate and spurious muons are now evaluated correctly. Before they were effectively switched off. Decisions of these filters are not used anymore to reject failing events but only to tag them for an offline analysis.

Other changes:

• As said above, events with duplicate or spurious muons are now tagged, and the corresponding flags are stored in an additional tree.
• Setup has been updated integrating a fix for MET recipe. This is not expected to change any stored quantities though.

Format of produced files has been changed in a backward-compatible way.

3.3.0-rc1

Candidate release to run on 03Feb2017 re-MiniAOD data and Summer16 simulation.

• Migrated to CMSSW_8_0_26_patch1.
• Global tags have been updated giving up-to-date JEC uncertainties.
• JER factors haven been updated to the latest version (Spring16_25nsV10), which is not final though.
• Script crabSubmit.py now tries to submit a task multiple times if the submission fails with a server error. Its output is now coloured.

Format of produced files has not been changed.

3.3.0-pre2

A tentative release for physics analysis with re-reconstructed 2016 data and Summer16 simulation.

Electrons have been reviewed and updated. Jets have been updated. This completes the object review for the updated setup. A summary of changes:

• Switched to preliminary JEC of version Summer16_23Sep2016_V2. They are read from local SQLite files. Uncertainties in this version are wrong.
• Added DeepCSV b-tagging discriminators.
• Cut-based and MVA electron ID have been updated to the 80X version, as well as electron isolation and trigger-emulating preselection.
• Added decisions of modified cut-based electron ID without the selection on isolation. They can be used in data-driven estimation of the multijet background. Only tight working point is considered.

Format of produced files is changed in a backward-compatible way.