TemplateForNewMCPPrescriptions

Latest recommendations for Release 21, based on MC16 and the full 13 TeV Run-2 dataset (2015-2018)

Previous Release 21 recommendations, now outdated, are described here: MCPAnalysisPreliminaryMC16, MCPAnalysisConsolidationMC16, MCPAnalysisWinter19MC16

Overview of latest prescriptions

The table below summarizes the set of relevant tools and configurations corresponding to the latest MCP recommendations. Further down in this page, more details and instructions are provided.

AnalysisBase>=21.2.100	TWiki	Tool	(Calibration)Release	Comments
Selection working points	MST	CP::MuonSelectionTool	-	WP: Loose, Medium, Tight, LowPt, HighPt
Reconstruction+selection efficiency	MESF	CP::MuonEfficiencyScaleFactors	191111_Winter_PrecisionZ	Use separate MESF tool instance, WorkingPoint = WP (see above)
Isolation efficiency	MESF	CP::MuonEfficiencyScaleFactors	191111_Winter_PrecisionZ	Use separate MESF tool instance, WorkingPoint = ISO (see here)
TTVA efficiency	MESF	CP::MuonEfficiencyScaleFactors	191111_Winter_PrecisionZ	Use separate MESF tool instance, WorkingPoint = TTVA
Momentum (q-agnostic)	MCAST	CP::MuonCalibrationAndSmearingTool	Recs2019_05_30	We suggest to use the MuonCalibrationPeriodTool.
Momentum (q-dependent)	MCAST	CP::MuonCalibrationAndSmearingTool	Recs2019_05_30	Please look at the detailed instructions here.

Useful references

A detailed description of the methodology used to extract the prescriptions and calibrations constants detailed below can be found at these links:

Link	Short description
ANA-MUON-2018-03-INT3	Internal note documenting the design and optimization of the muon selection levels for physics analysis.
ANA-MUON-2018-03-INT1	Internal note documenting the methods used to measure the reconstruction, identification and track-to-vertex association efficiencies, and the corresponding calibration factors.
ANA-MUON-2018-03-INT2	Internal note documenting the muon isolation selection criteria, and the measurement of the corresponding efficiencies and calibration factors.
coming soon ;-)	Internal note documenting the methods used to extract the muon momentum resolution and scale calibrations.

Release setup

Please use the (Ath)Analysis releases newer or equal to 21.2.100:

asetup AnalysisBase,21.2.100

What's new in 21.2.100:

• Reduction of systematics associated to the "bad muon veto" for the HighPt selection, and in general for pT>10 GeV.
• "Good" 2-station muons now included by default in the HighPt selection; can be turned off via a dedicated property of the selection tool.
• Revised high-eta scale factor maps with reduced systematic uncertainties.
• TTVA scale-factors with a new uncertainty covering non-closure in the crack.

Working points

Five muon identification selection levels are supported via the MuonSelectionTool (usage instructions are in the dedicated TWiki):

• Available selection WPs are :
  • Medium (default) : High efficiency down to pT~5 GeV, 0.1<|eta|<2.7; hadron misidentification rate at the level of few permille; small systematic uncertainties on SFs.
  • Loose : Maximal efficiency down to pT~3 GeV and for full eta coverage; hadron misidentification rate twice than Medium (especially at low pT), same for systematic uncertainties.
  • Tight : Efficiency from 5% to 10% lower than Medium; permille-level hadron misidentification rates; systematic uncertainties on SF larger than Medium.
  • LowPt : Maximal efficiency down to pT ~3 GeV; hadron misidentification rates and systematic uncertainties on SFs comparable to Medium.
    • Available option : UseMVALowPt property to turn on MVA for low-pT muons for testing purposes. Optimized to improve efficiency and hadron rejection.
  • HighPt : Best momentum resolution and removal of poorly measured high-pT tracks / "outliers", by requiring at least 3 MS stations and vetoing poorly aligned MS regions; the expected resolution roughly amounts to 12-15% for 1 TeV tracks.
    • Available option : Use2stationMuonsHighPt 2-station muons with missing inner MS station allowed for |eta|<1.3. Resolution loss on CB muons seems overall acceptable, feedback from analysis is welcome!

N.B. Any analysis selecting a non-negligible fraction of muons with pT>300 GeV should use the HighPt selection, or explicit approval from the MCP conveners will be required before it can proceed to publication. Please read carefully the guidelines detailed [[][here]] before deciding which selection level is most suitable to your case.

For analyses impacted by single events containing muon tracks with poorly measured momentum, a bad-muon veto functionality is also available, as detailed here.

Efficiency corrections

To get the reconstruction+selection, isolation, and track-to-vertex association (TTVA) efficiency scale factors, the MuonEfficiencyCorrections package is used. The corresponding tool is CP::MuonEfficiencyScaleFactors.

Please take a look at this page for further instructions on how to setup and use the tool correctly. Some usage examples can be found here:

• EventLoop: MuonEfficiencyScaleFactorsTest executable
• Athena: MuonEfficiencyCorrections_xAOD_Testing_jobOptions.py

Muon trigger efficiency scale factors

Trigger efficiency SFs are available for single and di-lepton triggers, and have been extracted for the full run 2 dataset: the relevant documentation, including usage instructions, can be found at this link. The default calibration area loaded in the tool is: 190129_Winter_r21.

Muon reconstruction efficiency scale factors

The muon reconstruction efficiency scale factors, and the corresponding systematics, are calibration factors and uncertainties that are meant to be applied on a per muon basis to events from MC16 simulation. They correct for (...)

Important notes

• The SFs are valid in the full pT down to 3 GeV. They are derived from J/Psi and Z events. The J/Psi measurement is used below 15 GeV. The threshold can be changed via the LowPtThreshold property, but this is not recommended.
• For Z-derived reconstruction+identification (pT>15 GeV) and TTVA SFs, it is now possible to retrieve the full breakdown of the Tag&Probe systematic uncertainties (instructions are below)
• The reconstruction efficiency and SFs for muons close to jets have been studied in detail using J/Ψ → μμ events, as documented here. SFs were found to be consistent with those for well-isolated muons, for different isolation and ΔR(jet,μ) criteria. No additional systematic uncertainty are needed.

How to setup a tool instance to retrieve reconstruction and selection efficiency scale factors

The tool can be configured via the Property mechanism. If your code uses the asg::AnaToolHandle then you can do it via properties for AnalysisBase and AthAnalysis. Otherwise use joboption files for AthAnalysis. For nominal operation, you only need to set the following properties:

This means muons have to be preselected using MuonSelectionTool (see above). To control the calibration release (not necessary for normal usage):

Option	Default value	Available values	Additional comments
CalibrationRelease	"191111_Winter_PrecisionZ"	Calibration releases available at calibration area	The calibration release to use (NB: not recommended to change! Please use the default in the tool).

Supported systematics

The CP::MuonEfficiencyScaleFactors tool supports the ASG systematic interface. The name of systematic option is as follows :

Name	Meaning	Types
MUON_EFF_RECO_STAT	Statitical error on the pT>15 GeV component of the SF	+/-1
MUON_EFF_RECO_SYS	Systematic error on the pT>15 GeV component of the SF	+/-1
MUON_EFF_RECO_STAT_LOWPT	Statitical error on the pT<15 GeV component of the SF	+/-1
MUON_EFF_RECO_SYS_LOWPT	Systematic error on the pT<15 GeV component of the SF	+/-1

Option for systematics breakdown

For Z-derived reconstruction and selection scale factors (pT>15 GeV), it is now possible to retrieve a full breakdown of the Tag&Probe systematic uncertainties, by setting to true the BreakDownSystematics property (default is false). NB: this is an option that should be tested by analyses sensitive to correlations or profiling of the corresponding SF systematics.

Bad muon veto (efficiency) systematics for High-pT selection

When the bad muon veto is applied and the HighPt selection WP is used, a dedicated systematic uncertainty needs to be propagated to MC. This is equivalent to an efficiency SF systematics, and has to be applied for each muon passing the veto. The corresponding SF is simply returned as 1 form the tool.

The systematics can be retrieved from an instance of CP::MuonEfficiencyScaleFactors by setting the property WorkingPoint to BadMuonVeto_HighPt. Then, use the usual mechanism to retrieve SF uncertainties via the ASG systematic interface.

Muon isolation efficiency scale factors

The list of recommended and supported isolation selection working points, together with a brief description, can be found under the IFF sub-group TWiki page at this link.

How to setup a tool instance to retrieve isolation efficiency scale factors

The scale factors and efficiencies can be obtained by passing the desired isolation working point name via the WorkingPoint property to an instance of the CP::MuonEfficiencyScaleFactors tool. Besides that, the configuration of the tool, including the recommended CalibrationRelease property, is completely analogous to the case of reconstruction and selection efficiencies.

Important notes

• A suffix "Iso" needs to be added to the isolation working point name.
• A prefix "LowPt_" is recommended for whoever is using the LowPt selection WP, and will retrieve tailored isolation SFs with reduced systematics.
• This need to be done using a separate CP::MuonEfficiencyScaleFactors instance, not the one used to retrieve the reconstruction SFs.
• Studies about efficiency and fake rejection on simulation are reported in this presentation
• NB: We suggest to investigate interaction between isolation WP and overlap-removal, if an analysis could be sensitive to it: preliminary studies are available here.
• SF are provided for muons with pt>3 GeV.
• SF for WPs using FixedRad (i.e floor of ptVar cone isolation) or with pTVar cone shrinking to zero are identical.
• For pT>500 GeV muons, the SF measured in 300-500 GeV region is provided, with a total systematic uncertainty 1% for trackOnly WP and 2% for other WP.

Supported systematics

The CP::MuonEfficiencyScaleFactors tool supports the ASG systematic interface. The name of systematic option is as follows:

Name	Meaning	Types
MUON_EFF_ISO_STAT	Statitical error on the SF	+/-1
MUON_EFF_ISO_SYS	Systematic error on the SF	+/-1

Muon track-to-vertex-association (TTVA) efficiency scale factors

Muon track-to-vertex association (TTVA) efficiency corrections are derived only for the standard tracking recommendations for muons, described at this link and consisting in: |d0BL significance | < 3 AND |Δ z0BL sin θ| < 0.5 mm. These corrections factorize with the other scale factors, and should be applied if you use these cuts in your analysis.

How to setup a tool instance to retrieve TTVA efficiency scale factors

The scale factors and efficiencies can be obtained by passing the working point TTVA to an instance of the CP::MuonEfficiencyScaleFactors tool. Apart from the working point, the configuration of the tool, including the recommended CalibrationRelease property, is completely analogous to the case of reconstruction and selection efficiencies.

Supported systematics

The CP::MuonEfficiencyScaleFactors tool supports the ASG systematics interface. The name of systematic option is as follows :

Name	Meaning	Types
MUON_EFF_TTVA_STAT	Statitical error on the SF	+/-1
MUON_EFF_TTVA_SYS	Systematic error on the SF	+/-1

Option for systematics breakdown

For TTVA scale factors, it is now possible to retrieve a full breakdown of the Tag&Probe systematic uncertainties, by setting to true the BreakDownSystematics property (default is false). NB: this is an option that should be tested by analyses sensitive to correlations or profiling of the corresponding SF systematics.

Momentum corrections

To apply momentum corrections to reconstructed muons, the MuonMomentumCorrections package is used. The corresponding tool is CP::MuonCalibrationAndSmearingTool. The tool must be applied to both data and MC separately for each data-taking year, and calibrates MC momentum scale and resolution to match data. Further, it allows to improve on data the residual momentum bias for mu+ vs mu- (sagitta bias correction) tracks.

The MuonCalibrationPeriodTool provides a wrapper-tool that we suggest to ease the application of the corrections to the entire set of Run-2 data and simulation samples.

Please take a look at this page for further instructions on how to setup and use the tool correctly. Some usage examples can be found here:

• EventLoop: MuonCalibrationAndSmearingTool executable

Momentum resolution and scale calibrations (charge-agnostic)

How to setup a tool instance to retrieve charge-agnostic calibrations

The CP::MuonCalibrationAndSmearingTool tool can be configured through the usual Property mechanism, for example:

Option	Default value	Supported values	Additional comments
Release	Recs2019_05_30	Calibration releases available at calibration area	The calibration release to use (NB: not recommended to change! Please use the default in the tool).

The default configuration is set to pickup the latest smearing and scale (charge blind) calibrations for the 2015-2016 dataset. Moreover, the default configuration is with sagitta bias correction turned OFF (see Setup2 below); feedback on its behavior is welcome!

N.B. Each instance of the tool needs to be configured separately, as the default instantiation is agnostic of the specific setup. In simulation, events need to be split proportionally among the different instances. For inclusive analyses over the full dataset the simplest way to do so is via the random run number from the PRW tool. An example on how to code the splitting can be found in the How to handle Monte Carlo section of the Muon Momentum Corrections Subgroup twiki. We however suggest to take advantage of the MuonCalibrationPeriodTool, unless a configuration different from what provided by default is desired.

Supported systematics

The CP::MuonCalibrationAndSmearingTool tool supports the ASG systematics interface. Here is a list of the independent systematics available, relevant to charge-agnostic momentum corrections:

Name	Meaning	Types
MUON_ID	Variations in the smearing of the ID track	+/-1
MUON_MS	Variations in the smearing of the MS track	+/-1
MUON_SCALE	Variations in the scale of the momentum	+/-1

Sagitta bias correction (charge-dependent)

How to setup a tool instance to retrieve charge-dependent calibrations

Depending on the analysis, we recommend two possible configurations, as listed below.

Setup 1 (correct data for sagitta bias)

This setup is meant for measurements dominated by muons of pT < 200 GeV and not statistically limited by events in the tail of the pT distribution. Data are corrected directly for the measured sagitta bias distortions, separately for the 2015, 2016, 2017 and 2018 campaigns. Please configure the tool as follows:

For 2015-2016 datasets:

For 2017 dataset:

For 2018 dataset:

Setup 2 (data untouched, additional systematic on MC for sagitta bias)

This setup is for measurements limited by statistics in the high tails of the pT distribution, or searches involving muons with pT >>200 GeV. In this case, we do not recommend the application of the sagitta bias correction on data not to be sensitive to local inhomogeneities of the sagitta bias maps. However, an uncertainty corresponding to the size of the bias is assigned to MC to cover for this effect. Please configure the tool as follows:

For 2015-2016 dataset:

For 2017 dataset:

For 2018 dataset:

Supported systematics

The CP::MuonCalibrationAndSmearingTool tool supports the ASG systematics interface. Here is a list of the independent systematics available, relevant to charge-dependent momentum corrections:

Name	Meaning	Types
MUON_SAGITTA_RHO	Variations in the scale of the momentum (charge-dependent), based on combination of correction on combined (Z scale) and recombination of the corrections	+/-1
MUON_SAGITTA_RESBIAS	Variations in the scale of the momentum (charge-dependent), based on the residual charge-dependent bias after correction	+/-1

Major updates:
-- StefanoZambito - 2020-01-27

%RESPONSIBLE% StefanoZambito
%REVIEW% StefanoZambito