Cutflow: pass trigger and at least 2 baseline leptons selection

[A-A-Abdelhamid]

I have been investigating why my acceptance values are off 0.5492 compared to the paper 0.682 in the 400 GeV tau= 1 ns sample by plotting distributions of pt, eta, etc... but nothing stood out as strange, so I looked into applying selection cut process itself by creating a cutflow.

I originally did NOT use a trigger selection criteria in any of the previous samples (including the the very first one that I worked on for a while with 400 GeV and proper tau of 0.1 ns)

With NO trigger selection cut, the acceptance is 0.4115, compared to the paper value of 0.386, which makes sense (relaxing a selection cut condition corresponds to more acceptance)

Let's apply the trigger selection cut:

I generated 20,000 events with mass of 500 GeV and proper lifetime of 0.1 ns

By applying the first selection using this "trigger" function:

def triggered(event):
    has_valid_electron = False
    has_valid_muon = False
    electron_count = 0
    lepton_eta_count = 0  # Count of leptons with eta < 2.5

    for particle in event.particles:
        if particle.status != 1:  # Consider only final state particles
          continue

        eta = CalcEta(particle)
        if abs(particle.pid) == 11:  # Electron
            if particle.momentum.pt() > 160:
                has_valid_electron = True
            elif particle.momentum.pt() > 60:
                electron_count += 1

        elif abs(particle.pid) == 13:  # Muon
            if particle.momentum.pt() > 60 and abs(eta) < 1.05:
                has_valid_muon = True

        if abs(particle.pid) in [11, 13] and abs(eta) < 2.5:
            lepton_eta_count += 1

    # Check if the event satisfies the original conditions
    if has_valid_electron or electron_count >= 2 or has_valid_muon:
        # If original conditions are met, check for the additional eta condition
        if lepton_eta_count >= 2:
            return True

    return False

And using it as follows:

for event in f:

***************
***************
      cond=triggered(event)
      if cond == False:
        nottriggered = nottriggered +1
        continue

      for particle in event.particles:
**************
*************

And here is what I got:

Fraction of events that passed in the generated sample: 0.8277
Error_Low: 0.0027118284088818445
Error_high: 0.0026787525426837933
Fraction of events that passed in the paper: 0.7083      #(66.3 out of 93.6)
Error_Low: 0.0545094960581447
Error_high: 0.04958628972360113

The "error" here is the statical uncertainty assuming a Poisson distribution

In my sample the fraction events that passed the trigger and 2 baseline leptons cut is 0.8277 +/- 0.0027 while it is 0.7083 +0.04956/-0.0545 in the paper

The sample acceptance after applying all the selection cuts (after the already applied trigger cut) is 0.3599 ~ 0.360, compared to 0.386 in the paper. With statistical error of +/- 0.0034, the acceptance has an upper estimate of 0.363, compared to the paper 0.386

By applying the trigger condition in the selection process, acceptance went from 0.412 to 0.360 which agrees more with the paper. I need to go back and impose this trigger condition to get a more accurate acceptance, yield, and expected number of events. The main sample of 400 GeV and tau = 0.1 ns has a slightly higher acceptance than the paper, so I expect this to be fixed after applying the trigger cut. However, I am still looking into my 400 GeV and tau = 1 ns sample, as its acceptance is lower than the paper. Although the last sample (400 GeV and tau = 1 ns) is what provoked me to look into the cutflow and triggering conditions, by applying the trigger selection cut the acceptance can only go down creating more discrepancy.

Next steps:

So my next steps is to:

1) complete the cutflow to make sure there is no more discrepancy between what we are doing and what the paper did 2) Continue looking into the 400GeV tau = 1ns sample