ABCDEFGHIJKLMNOPQRSTUV
1
JEC NameShort Description from Run2 twikiCorrelated CMS-ATLAS (Run1)?Why / Motivation (Run1 CMS-ATLAS) from JME-15-001Correlated CMS-ATLAS (Run2)?Why / Motivation
2
AbsoluteStatStatistical uncertainty on in situ calibration with photon-jet/Z-jet events0% (uncorrelated) - statistical in situ termGroup 1 'Statistical and detector-based uncertainties from in situ techniques': ATLAS and CMS use statistically distinct data and MC samples and different detector technologies, so these are uncorrelated.0% (uncorrelated) - statistical in situ term-> still for Run2KEEP SAME MOTIVATION AS: JME-15-001Group 1 'Statistical and detector-based uncertainties from in situ techniques': ATLAS and CMS use statistically distinct data and MC samples and different detector technologies, so these are uncorrelated.
3
AbsoluteScaleFlat absolute scale uncertainty (combined photon, Z->ee, Z->mumu reference scale)0% (uncorrelated) - detector in situ term Group 1 (detector-based, Table 4 row 1b): detector-description-driven terms are treated the same as the statistical ones - fixed at 0% since they rely on each experiment's own detector calibration.0% (uncorrelated) - detector in situ term -> still for Run2KEEP SAME MOTIVATION AS: JME-15-001Group 1 (detector-based, Table 4 row 1b): detector-description-driven terms are treated the same as the statistical ones - fixed at 0% since they rely on each experiment's own detector calibration.
4
AbsoluteMPFBiasMPF method bias correction (neutrinos and ISR+FSR outside detector acceptance)0-50% (partially correlated) - absolute balance modeling termGroup 2 'Modeling uncertainties from in situ techniques': ATLAS and CMS measure similar effects (neutrinos in jets, out-of-cone, radiation) but combine them differently (ATLAS: pT-binned combination; CMS: global 2-parameter fit) -> partially correlated, scanned 0-50%.0-50% (partially correlated) - absolute balance modeling term -> still for Run2KEEP SAME MOTIVATION AS: JME-15-001 Group 2 'Modeling uncertainties from in situ techniques': ATLAS and CMS measure similar effects (neutrinos in jets, out-of-cone, radiation) but combine them differently (ATLAS: pT-binned combination; CMS: global 2-parameter fit) -> partially correlated, scanned 0-50%.
5
FragmentationHigh-pT extrapolation based on Pythia6 Z2 / Herwig++2.3 differences in fragmentation and underlying event (FullSim). Which models have been used? Full uncorellated can be conservative0% (uncorrelated) - grouped with High-pT uncertainties. Group 8 'High-pT uncertainties': the CMS Fragmentation term is mapped (Table 4, row 8) to ATLAS' High-pT uncertainty; methodology and test-beam energies used differ between experiments, so treated as uncorrelated.?To be discussed with ATLAS. Did the ATLAS method change? Which tune they use (e.g. Pythia8)?
6
SinglePionECALHigh-pT extrapolation based on propagation of ±3% variation in single particle response in ECAL to PF jets (FastSim)0% (uncorrelated) - detector in situ term Group 1b 'Detector in situ terms': single-hadron response measurements are experiment-specific (different test-beam/detector data), fixed at 0%.0% (uncorrelated) - detector in situ term -> still for Run2KEEP SAME MOTIVATION AS: JME-15-001 Group 1b 'Detector in situ terms': single-hadron response measurements are experiment-specific (different test-beam/detector data), fixed at 0%.
7
SinglePionHCALHigh-pT extrapolation based on propagation of ±3% variation in single particle response in HCAL to PF jets (FastSim)0% (uncorrelated) - statistical in situ term Group 1a 'Statistical in situ terms': in Run1 this component is constrained by the CMS global fit and becomes a statistical term, so grouped with the purely statistical/detector uncertainties, fixed at 0%.0% (uncorrelated) - statistical in situ term -> still for Run2KEEP SAME MOTIVATION AS: JME-15-001 Group 1a 'Statistical in situ terms': in Run1 this component is constrained by the CMS global fit and becomes a statistical term, so grouped with the purely statistical/detector uncertainties, fixed at 0%.
8
FlavorQCDDefault jet-flavor uncertainty, based on Pythia6 Z2/Herwig++2.3 differences in uds/c/b-quark and gluon responsesNot a single value - Run1 note splits flavor by parton type (see FlavorPureGluon/Quark/Charm/Bottom below). For Run2: we should consider the correlation? We need info from ATLAS for Run2Section 4 groups 4-6: the note does not treat 'flavor' as one component; it separates gluon-jet (group 4), b-jet (group 5) and other quark/charm flavors (group 6), each with its own correlation, because the modeling robustness and the ATLAS/CMS derivation methods differ by parton type.?To be discussed with ATLAS. Is also ATLAS considering a gloabal source of uncertainty for FlavourQCD? The low response observed in Run 1 is expected to have a common component between CMS and ATLAS, while the underlying event has a relatively small impact.
9
FlavorPureGluonAlternative flavor uncertainty source, gluon-only jets100% (fully correlated)Group 4 'g-jet fragmentation energy scale uncertainty': both experiments derive the uncertainty from the same PYTHIA/HERWIG++ difference; Figure 1 in the note shows the resulting ATLAS and CMS uncertainties are nearly identical across pT and |eta|<3.2, so the components are taken as fully correlated.100% (fully correlated)->still in Run2KEEP SAME MOTIVATION AS: JME-15-001 Group 4 'g-jet fragmentation energy scale uncertainty': both experiments derive the uncertainty from the same PYTHIA/HERWIG++ difference; Figure 1 in the note shows the resulting ATLAS and CMS uncertainties are nearly identical across pT and |eta|<3.2, so the components are taken as fully correlated.
10
FlavorPureQuarkAlternative flavor uncertainty source, uds-only jets0% (uncorrelated) - other fragmentation typesGroup 6 'Other jet fragmentation energy scale uncertainties': ATLAS folds light-quark differences into its larger gluon-uncertainty budget, while CMS derives a dedicated per-flavor PYTHIA/HERWIG++ comparison; the note states the two techniques are 'unrelated... and not directly comparable', hence uncorrelated.0% (uncorrelated) - other fragmentation types->still in Run2KEEP SAME MOTIVATION AS: JME-15-001 Group 6 'Other jet fragmentation energy scale uncertainties': ATLAS folds light-quark differences into its larger gluon-uncertainty budget, while CMS derives a dedicated per-flavor PYTHIA/HERWIG++ comparison; the note states the two techniques are 'unrelated... and not directly comparable', hence uncorrelated.
11
FlavorPureCharmAlternative flavor uncertainty source, charm-only jets0% (uncorrelated) - other fragmentation typesGroup 6, same reasoning as FlavorPureQuark: ATLAS and CMS derivation methods for non-gluon, non-b flavors are not directly comparable, so treated as uncorrelated.0% (uncorrelated) - other fragmentation types -> still in Run2KEEP SAME MOTIVATION AS: JME-15-001 Group 6, same reasoning as FlavorPureQuark: ATLAS and CMS derivation methods for non-gluon, non-b flavors are not directly comparable, so treated as uncorrelated.
12
FlavorPureBottomAlternative flavor uncertainty source, bottom-only jets50-100% (mostly correlated)Group 5 'b-jet fragmentation energy scale uncertainty': both derive the uncertainty from PYTHIA/HERWIG++ differences with substantial methodological similarity, but ATLAS parametrizes it vs pT (flat in eta) while CMS does the opposite, adding some decorrelation - hence mostly correlated, scanned 50-100%.50-100% (mostly correlated) -> still in Run2KEEP SAME MOTIVATION AS: JME-15-001 Group 5 'b-jet fragmentation energy scale uncertainty': both derive the uncertainty from PYTHIA/HERWIG++ differences with substantial methodological similarity, but ATLAS parametrizes it vs pT (flat in eta) while CMS does the opposite, adding some decorrelation - hence mostly correlated, scanned 50-100%.
13
TimePtEtaJEC time dependence between data-taking epochs (BCD, EF, G, H); L2L3Residual now corrected per epoch. Full 2016 sample, the uncertainty is not necessary.0% (uncorrelated) - single-experiment termGroup 9 'Other uncertainties': a small number of remaining sources evaluated in one experiment but not the other are taken as uncorrelated between ATLAS and CMS; Table 4 explicitly maps TimeEta/TimePt here. Does atlas have this uncertainty?0% (uncorrelated) - single-experiment term-> still in Run2KEEP SAME MOTIVATION AS: JME-15-001 Group 9 'Other uncertainties': a small number of remaining sources evaluated in one experiment but not the other are taken as uncorrelated between ATLAS and CMS; Table 4 explicitly maps TimeEta/TimePt here. Does atlas have this uncertainty?
14
RelativeJER[EC1][EC2][HF]Eta-dependence uncertainty from jet pT resolution (JER), derived by varying JER scale factors up/down0% (uncorrelated) - detector in situ term -> still in Run2. ATLAS is using the same method of Run1?Group 1b 'Detector in situ terms' (Table 4): JER data/MC differences depend on each experiment's own detector resolution measurement, fixed at 0%.0% (uncorrelated) - detector in situ term->still in Run2. To be checkedTo be discussed with ATLAS. Does ATLAS use the same method of Run1 (to be checked)?
15
RelativePt[BB][EC1][EC2][HF]Half-difference between MPF log-linear (default) and constant fits vs pT0% (uncorrelated) - detector in situ term->still in Run2. What atlas uses for the parametrization?Group 1b 'Detector in situ terms' (Table 4): treated alongside AbsoluteScale/SinglePionECAL/RelativeJER as a detector-driven, experiment-specific term, fixed at 0%.0% (uncorrelated) - detector in situ term->still in Run2. To be checkedTo be discussed with ATLAS. What ATLAS used for the parametrisation uncertainty?
16
RelativeBalFull difference between log-linear fits of the dijet MPF and pT balance methodsN/A - source does not exist in the Run1 note, ATLAS introduces something equivalent?N/A - RelativeBal is not defined anywhere in this note (Tables 3/4 or Section 3); it was introduced only for Run2, so no Run1 correlation can be quoted from this source.?To be discussed with ATLAS. Did ATLAS introduce an equivalent uncertainty?
17
RelativeSampleEta-dependent uncertainty due to the difference between relative residuals observed with dijet, Z+jets and gamma+jets samplesN/A - source does not exist in the Run1 note, ATLAS introduces something equivalent?N/A - same as RelativeBal: not present in this note's Table 3/4, introduced only in Run2.?To be discussed with ATLAS. Did ATLAS introduce an equivalent uncertainty?
18
RelativeFSREta-dependence uncertainty due to correction for initial and final state radiation50-100% (mostly correlated) Group 3 'Relative eta-intercalibration modeling uncertainty': both experiments use a generator-difference (PYTHIA vs HERWIG++ for CMS; similar approach for ATLAS) to model radiation in the dijet eta-intercalibration; given the similar generators/techniques, treated as mostly correlated, scanned 50-100%.Decorellate? or partly correlate (0-50%)?To be discussed with ATLAS.
19
RelativeStat[FSR][EC][HF]Statistical uncertainty in the determination of eta-dependence (FSR correction fit / L2Res log-linear fit)0% (uncorrelated) - statistical in situ term- -> still in Run2Group 1a 'Statistical in situ terms' (Table 4): purely statistical uncertainties from distinct data samples/detectors, fixed at 0%.0% (uncorrelated) - statistical in situ term- -> still in Run2KEEP SAME MOTIVATION AS: JME-15-001 Group 1a 'Statistical in situ terms' (Table 4): purely statistical uncertainties from distinct data samples/detectors, fixed at 0%.
20
PileUpDataMC5% uncertainty on the data/MC scale factor for offset correction (roughly covers mu variation seen with Random Cone on Zero Bias data / neutrino gun MC)0% (uncorrelated) - pileup term->still in Run2Group 7 'Pileup uncertainties': although both use variants of FastJet area-based pileup subtraction, the actual derivation of the uncertainty differs dramatically (CMS: pT-dependent offset difference in/out of jets; ATLAS: 4 separate NPV/mu/rho-based terms), so treated as uncorrelated.0% (uncorrelated) - pileup term->still in Run2KEEP SAME MOTIVATION AS: JME-15-001 Group 7 'Pileup uncertainties': although both use variants of FastJet area-based pileup subtraction, the actual derivation of the uncertainty differs dramatically (CMS: pT-dependent offset difference in/out of jets; ATLAS: 4 separate NPV/mu/rho-based terms), so treated as uncorrelated.
21
PileUpPt[Ref][BB][EC1][EC2][HF]Pile-up offset dependence on jet pT, estimated from matched MC with/without PU overlay vs Random Cone method0% (uncorrelated) - pileup term->still in Run2Group 7, same reasoning as PileUpDataMC: pileup suppression is conceptually similar but the uncertainty-derivation methods differ substantially between the two experiments, so uncorrelated.0% (uncorrelated) - pileup term->still in Run2KEEP SAME MOTIVATION AS: JME-15-001 Group 7, same reasoning as PileUpDataMC: pileup suppression is conceptually similar but the uncertainty-derivation methods differ substantially between the two experiments, so uncorrelated.
22
JERN/A in the note?To be discussed with ATLAS.
23
24
25
26
27
28