Notes on Tracking

• Track Parameters derived from base pricinicles

Tracking performance study with Boris

Boris's tasks

Task 1

Task: Learn how to create a root tree with simple track variables.

Status: Complete

• Screenshot of myTrackVariables:
  
  

Task 2

Task: Learn how to plot the simple variables in the root tree using TBrowser and the mouse. Do simple histogram manipulations as fitting, re-binning and adding labels from the graphical interface.

Status: Complete

• Using TBrowser and graphical fit options:
  
  

Task 3

Task: Repeat Task 2 using a root macro

• Open your root file which contains the root tree.
• Access the root tree.
• Plot one simple variable (in this case z_{0}).
• Fit the histogram.
  
• Get the fit parameters back from the histogram object. You should be able to write these to a file if necessary.

Status: Complete

• Using a macro to plot and fit z_{0}:
  
  

• Exercise in root fitting.:
  
  

Task 4

Task: Add additional variables to the root tree related to the simTracks. Use the same variables as reco::track variable tree.

Status: Incomplete

Task 5

Task: Learn how to associate a reco::Track to a simTrack and vice versa. After you learn that, we can add to root tree some more interesting numbers like the residual of track parameters:

• deltaTIP: recotrack_transverseImpactParameter - simTrack_transverseImpactParamer
• Δp_T/p_T
• Δφ
• Δη
• fraction of collected rechits with respect to the simulated ones.
• purity of the collected rechits
• track reconstruction effeciency and fate rate flag
• etc.

Status: Incomplete

Task 6

Task: Plot all the values stored in the tree at point 5 versus other simtrack and reco::track parameters

• efficiency vs η (you can compare with official plot)
• efficiency vs p_T
• resolution of p_T vs purity of the collected hits (another new plot)
• etc.

Status: Incomplete

Task 7

Task: After we have arranged all the previous machinery for a single track sample, we will move to a real multitrack sample to see what happens to track reconstruction inside a collimated jet or for a very energetic μ.

Status: Incomplete

Plot's for Prof. Sharma

For a sample of 50,000 Z→μμ events:

• d0 of Tracks with P_T > 15 GeV and χ²/d.o.f. < 4:
  

• d0 of Tracks with P_T > 15 GeV and χ²/d.o.f. ≥ 4:
  

• z0 of Tracks with P_T > 15 GeV and χ²/d.o.f. < 4:
  

• z0 of Tracks with P_T > 15 GeV and χ²/d.o.f. ≥ 4:
  

Plot for Boris

• This is from 50k Z→μμ events with P_T > 15 GeV and χ²/d.o.f. < 4 and number of valid hits > 8:
  
  

Track Quality Study

Various kinematical variables of reconstructed tracks Tracks.

No Cuts (RecoTrack)

I loop over all reconstructed tracks in an event and plot the following kinematical variables and detector information: P_T, η, φ, d₀, z₀, and the number of valid hits in the tracker.

• ~50k Z→μμ events. (pdf version)
  
  

• ~120k QCD (50 < P_T < 80) events. RawVariablesQCD50to80.pdf

Matched vs Unmatched.

In each event I have two loops:

• The first loop is over the RecoTracks in each event. Next I define a match as a RecoTrack that has a corresponding SimTrack associated by hits using CMSSW_1_3_1 TrackAssociatorByHits that Boris showed me. Unmatched is defined as NOT Matched.
  • 50,000 Z→μμ events: (pdf version)
    
    

• The Second loop is over the SimTracks in each event. Next I define a match as a SimTrack that has a corresponding RecoTrack associated by hits again using CMSSW_1_3_1 TrackAssociatorByHits. Unmatched is defined as NOT Matched.
  • 50,000 Z→μμ events: (pdf version)
    
    

With Cuts

• The following cuts were applied:
  
  • |η| < 2.5
  • # valid hits > 7
  • |d₀| < 1 mm
  • |z₀| < 30 cm
  • χ²/d.o.f. < 10

• RecoTracks that have an associated SimTrack.
  • 50,000 Z→μμ events: cutsMatchedReco.pdf
    
    

• RecoTracks that do NOT have an associated SimTrack.
  • 50,000 Z→μμ events: cutsNotMatchedReco.pdf
    
    

Δ = Reco - Sim.

• Calculated the difference between RecoTrack and SimTrack variables.
  • 50k Z→μμ events: Delta.pdf
    
    

Pull = ΔVariable / VariableError. Note P_T error has not been implemented yet and I'm not sure that Δφ has been calculated correctly.

• Pull of Various Kinematical Variables:

-- RyanKelley - 01 Aug 2007

• RawVariables.pdf: Various kinematical variables or reco Tracks (no Cuts).

• RawVariablesQCD50to80.pdf: RawVariablesQCD50to80? .pdf