6
C.Clément Tile commissioning meeting – 07-01-18 From susy group talk of last Wednesday Simulation and digitization is done in version 12.0.31.7 (8) Reconstruction 12.0.4 is failure -> Should use 12.0.5 12.0.5 ready (validated) within 1-2 weeks Opportunity to validate the 12.0.5 recoed samples MC samples available in ~1month Here I looked at the status of EventView in 12.0.4 Software Versions 1

C.ClémentTile commissioning meeting – 07-01-18 From susy group talk of last Wednesday Simulation and digitization is done in version 12.0.31.7 (8)

Embed Size (px)

Citation preview

Page 1: C.ClémentTile commissioning meeting – 07-01-18 From susy group talk of last Wednesday  Simulation and digitization is done in version 12.0.31.7 (8)

C.Clément Tile commissioning meeting – 07-01-18

From susy group talk of last Wednesday

Simulation and digitization is done in version 12.0.31.7 (8)

Reconstruction 12.0.4 is failure -> Should use 12.0.5

12.0.5 ready (validated) within 1-2 weeks Opportunity to validate the 12.0.5 recoed samples MC samples available in ~1month

Here I looked at the status of EventView in 12.0.4

Software Versions

1

Page 2: C.ClémentTile commissioning meeting – 07-01-18 From susy group talk of last Wednesday  Simulation and digitization is done in version 12.0.31.7 (8)

C.Clément Tile commissioning meeting – 07-01-18

How could cut flow look like?1. emu or ee (tightest lepton criteria applied)2. + 1 jet3. + 2 jets

1. Question: how many jets needed for SUSY dilepton analysis?2. Plot Njet for Z+jets (with the various njet bins properly weighted and SUSY

points),

4. ETmiss cut5. Other SUSY topology cuts?

Overlap removal

2

Types of leptons = loose⊃ medium⊃ tight

Eg. Loose: an electron candidate from the egamma group (El_author==1)This is very loose (ala D0, I am not sure this is the best)

Eg. Tight = Loose + isEM

Remove a jet if it overlaps with a Loose? Medium? Tight electron?

Page 3: C.ClémentTile commissioning meeting – 07-01-18 From susy group talk of last Wednesday  Simulation and digitization is done in version 12.0.31.7 (8)

C.Clément Tile commissioning meeting – 07-01-18

Overlap removal (2)

3

1) Remove jet if it overlaps with a Loose electron.

Many (real) jets will look like a loose electron, so many events will have jets removed, with the risk to decrease Njet for background, the risk is smaller for the signal.

Safe side from the electron efficiency point of view since we favour higher electron efficiency, we prefer to consider event loose EM objects as electrons than as jets

From S/B point of view, dilepton signature much more powerful tool to decrease QCD than Njet (reasonable guess), so again even though we decrease Njet by mistake, that should not lose us as much discrimination against QCD as decreasing nlepton

2) Remove only if it overlaps with a Tight electron?

Problem with fake background estimation, because Njet might not be the same in the loose and tight lepton samples

Page 4: C.ClémentTile commissioning meeting – 07-01-18 From susy group talk of last Wednesday  Simulation and digitization is done in version 12.0.31.7 (8)

C.Clément Tile commissioning meeting – 07-01-18

Overlap removal for reconstructed objects Overlap removal and insertion criteria for truth information: did not look

at that yet.

Particles are inserted in the EventView using ”Inserters” for each particle type (Atlas/PhysicsAnalysis/EventViewBuilder/EventViewInserters/)

Insertion order is steered up in EventViewBuilder/EventViewBuilderAlgs/share/defaultEventView_jobOptions.py

Overlap removal (3)

4

Order matters

pT ordered

The first inserted electron is the highest pT one found in theelectron container

Page 5: C.ClémentTile commissioning meeting – 07-01-18 From susy group talk of last Wednesday  Simulation and digitization is done in version 12.0.31.7 (8)

C.Clément Tile commissioning meeting – 07-01-18

Electrons are inserted in order of decreasing pT

Electron_i is inserted only if its DR to all electrons_j<i is larger than deltaRCut =0.1 (default)

Default defined in EventViewBuilder/EventViewInserters/src/EVElectronInserter.cxx

Can be changed in: defaultEventView_jobOptions.py

Some electron precuts put there as well, to filter what is going into the ntuple.

Overlap removal (4)

5

Jets are inserted in order of decreasing pT

Can decide also in defaultEventView_jobOption.py if special care has to be taken about overlap with

Other jets Tracks Calo clusters Muon

A certain DR can be defined as well (did not find default value yet)

Overlap cannot be removed with objects not yet inserted.

Page 6: C.ClémentTile commissioning meeting – 07-01-18 From susy group talk of last Wednesday  Simulation and digitization is done in version 12.0.31.7 (8)

C.Clément Tile commissioning meeting – 07-01-18

Find out more about the electron / jet removal, what is the defaut value of DR

The electron objects in the analysis should be matched to the electron object(s) in the trigger

Need to think about the trigger, because it will decide a lot of things

Study performance of ETmiss+jet Single lepton (+jet) Dilepton trigger

What to use for the first year of data taking?

next

6