UCSD vs UCSD 04072009 1 UCSD vs EWK

UCSD vs UCSD 04/07/2009 1

UCSD vs EWK 04/07/2009 2

Iso vs ID (barrel - 24_08_2009) Plot normalized to cross-section (signal in red x 10000) 04/07/2009 3

Iso vs ID (endcap - 24_08_2009) Plot normalized to cross-section (signal in red x 10000) 04/07/2009 4

Eff. Zee (24/08/2009) 04/07/2009 5

Eff. Zee (16/08/2009) 04/07/2009 6

Eff. Zee (with iso et dependence) 04/07/2009 7

Eff. Higgs (24/08/2009) 04/07/2009 8

Eff. Higgs (16/08/2009) 04/07/2009 9

Eff. Higgs (with iso et dependence) 04/07/2009 10

Fake rate (24/08/2009) 04/07/2009 11

Fake rate (16/08/2009) 04/07/2009 12

Fake rate (den – Very. Loose ele. ) 04/07/2009 13

Fake rate (den – Very. Loose ele. ) 04/07/2009 14

Iso vs ID (Barrel) Plot normalized to cross-section (signal in red x 10000) Iso 04/07/2009 Id 15

Iso vs ID (Endcap) Iso 04/07/2009 Id 16

Eff. Higgs (opt. including b/c jets) 04/07/2009 17

Eff. Zee (opt. including b/c jets) 04/07/2009 18

Fake rate (opt. including b/c jets) 04/07/2009 19

Fake rate/Jet (with iso et dependence) 04/07/2009 20

Lepton deposit removal. Different number of digits in the cuts. 04/07/2009 21

Fake Rate (dijets 0 -120 Ge. V) fake_rate per Gen. Jet fake_rate per Jet 04/07/2009 22

Fake Rate (QCD en 20 -80 Ge. V) fake_rate per Gen. Jet fake_rate per Jet 04/07/2009 23

Fake Rate (dy-> >20 Ge. V) fake_rate per Gen. Jet fake_rate per Jet 04/07/2009 24

Fake Rate (dy 10 -40 Ge. V) fake_rate per Gen. Jet fake_rate per Jet 04/07/2009 25

Gen. Jet vs Jet (DY-> ) - Gen. Jets - Reco Jets 04/07/2009 26

Eel/Egenjet (dy-> , barrel) Gen. Jet 04/07/2009 27 5 < ET < 10 10< ET< 20 ET > 20

Eel/Egenjet (dy-> , endcap) Gen. Jet 04/07/2009 28 5 < ET < 10 10< ET< 20 ET > 20

Total Eff. (Zee) Barrel Endcap 04/07/2009 29

Total Eff. (Higgs) Barrel Endcap 04/07/2009 30

Total Fake Rate (DY - QCD) Barrel Endcap 04/07/2009 DY QCD 31

“Efficiency” (BCto. E_20_80) Barrel p. T distribution of e passing e. ID selections: Zee bc->e (20_30) bc->e (30_80) Endcap Cut at generator level: e. TThreshold = cms. double(10. 0) 04/07/2009 32

Eff. vs ET (Zee - barrel) August results Last results 04/07/2009 33

Eff. vs ET (Zee - endcap) August results Last results 04/07/2009 34

Eff. vs ET (Higgs - barrel) August results Last results 04/07/2009 35

Eff. vs ET (Higgs - endcap) August results Last results 04/07/2009 36

Fake Rate per SC (QCD_20_80) Barrel Endcap 04/07/2009 37

Fake Rate per SC (DY_10_40) Barrel Endcap 04/07/2009 38

Zee OLD results with 1_6_9 Ol. D selection on the new samples 04/07/2009 39

QCD e-enriching Filter Between 1_6_7 and 2_X samples: at least one bug-fix: 04/07/2009 only e+ processed. improved the logic of the filter. 40

Higgs OLD results with 1_6_9 Ol. D selection on the new samples 04/07/2009 41

(Very) New Categorization Category re-defined as: barrel: endcap: fbrem < 0. 12 0. 9 < E/p < 1. 2 fbrem < 0. 2 0. 82 < E/p < 1. 22 Removed squared cut in E/p, f. Brem: New cuts: f. Brem > -2 Eseed/pin → Eseed/pin + f. Brem 04/07/2009 42

New Categorization (p. T > 20) 04/07/2009 43

New Categorization (10 < p. T < 20) 04/07/2009 44

New Categorization (5<p. T<10) 04/07/2009 45

Class Based e. ID (barrel) 04/07/2009 46

Class Based e. ID (endcap) 04/07/2009 47

EWK selection Full list of applied cuts: HLT_ele 15_LW_L 1 R ET(SC) > 30 Ge. V Excluded: 1. 4442 < | 1. 56 e. ID same variables as Robust: 04/07/2009 In endcap added also E/p and d 0. Usually done together tk, ecal and hcal iso. 48

Likelihood e. ID Processed with PAT: Zee and QCD. Apply “similar” cuts to compare with our selection. Looking for likelihood thresholds to have same signal efficiencies: 04/07/2009 49

Globe 2 This is only an exercise to see how PAT work and if is possible to “convert” globe: Implemented few classes to read PATuples. Make it more OO. Replacement of ROOT I/O with get. By. Label statements. Keep similar structure to globe to re-use as much as possible Marco's code. Probably not the best to do (a real analysis could be more probing). Also need to learn some other functionality of 04/07/2009 PAT. 50

Globe 2 Looper Class equivalent to Loop. All, read inputfiles. dat and take care of normalization number of events. . . Base. Analysis implement basic functionality needed by every analysis: Counters, Histograms, Cuts, csv, . . . Electron. Analysis Zee. Analysis Higgs. Analysis General. Functions 04/07/2009 51

Using PAT and Globe 2 Pros: Skip reduce step, Re-use of common functions: matching tools, isolation. . . Use of FWLite -> full CMSSW capabilities, Stand-alone C++ executable, easier to debug using gdb, Possibility to port directly the code in CMSSW (not sure 100%). 04/07/2009 52

Using PAT and Globe 2 Cons: Bigger “ntuple” something like a factor 5, Slightly slower: The quanta of information are bigger (branches not with simple floats but entire Data. Formats), CMSSW needed in the laptop, Necessity to run on samples multiple times with CRAB, 04/07/2009 53