Meeting minutes

[ytatus94]

• 20170412
  • Truth level study
  • What is this?
  • Mike asked: Is it pure Higgsino?
  • Judita answered: NUHM2 is a realistic model
  • We didn't generate pure compressed state because it has highest cross-section.
  • Signal jet + 2/3L

[ytatus94]

• 20170519
  • Use compressed + accessible or compressed + compressed
  • Do truth level study ==> Need to generate TRUTH3 derivation
  • Check the sensitivity for compressed + accessible
  • Produce the significance plots Zn
  • Request compressed + compressed production.
  • N1 < C1 < N2
  • How to produce mass spectrum?
  • Same as simplified model (Jesse Liu)
    • (N2-N1) / (C2-N1) should be two for simplified model
    • The (N2-N1) / (C2-N1) for NUHM2 varies from 1.6 (low m1/2) to 1.2 (high m1/2)
  • k-Factor should be 1.
  • Use log files to calculate 1 lepton filter efficiency
  • Shall we remove lepton filter?
  • Tommaso suggests 2L3 filter (3 means pT > 3 GeV).

[ytatus94]

• 20170522
  • Mike:
  • How long it will take to produce compressed + compressed?
  • If we use compressed + accessible, then we can look at the compressed side only.
  • Judita:
  • Do we need MadSpin?
  • Only generate m1/2 = 400, 600, 800 for compressed + compressed
  • Use the truth level study
  • HIGGSINO-24 JIRA ticket
  • Use TRUTH3 check m1/2 = 600 and compare with Higgsino_160_150
  • Use SimpleAnalysis to produce plots
  • EVNT ==> TRUTH3 ==> SimpleAnalysis ==> ntuple
  • EVNT ==> RECO ==> SUSY16 ==> framework to produce ntuple
  • Joey
  • b-jets veto uses nBjet30_MV2c10 == 0
  • Because we veto b-jets, we don't save b-jet pT information in the ntuple.
  • nTotalJets includes forward jets but nJet30 and nJet25 only include signal jets
  • jetPt also includes signal jets only
  • Why do we choose Higgsino_160_150?
  • Compare the mass spectrum and read the N1 and N2 value from the spectrum.
  • Try to find the Higgsino sample to choose one sample closed to N1 and N2 we read from the spectrum.
  • The existing Higgsino sample are quite useful for us
  • The mass splitting for Higgsino sample is the same between N2 - N1
  • Loss mass Z^* comes from N2.

[ytatus94]

• 20170601
  • Judita:
  • We current use compressed + accessible
  • We try to produce compressed + compressed
  • In between solution:
    • Do truth level study and compare with Higgsio sample
    • Scaling, make exclusion plots as Higgsion signal points
    • Update everything after we get compressed + compressed MC.
  • Check the filter efficiency in the log file.
  • Pat:
  • Meeting with Judita at 2:00pm at R1 on Friday.

[ytatus94]

• 20170602
  • EWK meeting today
  • Julia will present the status of Higgsino study includeing NUHM2
  • Generate the MC
  • compressed + compressed without filter
  • compressed + compressed + one jet without filter
  • compressed + accessible without filter
  • The filter efficiency is 0.3, so we have to generate 3 times larger number of events if we don't require filter in the new MC production
  • We don't require new mass points.
  • We want to get compressed + compressed MC as soon as possible.
  • Higher priority on m12=400, 600, 800
  • Envelope calculation:
  • Calculate N = sigma * lumi
  • Reweight mll and mT2 as Wino/Bino did.
  • If the Z is included in the decay, the Pythia cannot produce the decay correctly. We have to use Madspin for the compressed.
  • Madgraph --> Madspin --> Pythia

[ytatus94]

• 20170607
  • All requests put in one:
    • compressed + compressed
    • compressed + compressed + 1 jet
    • compressed + accessible
    • compressed + accessible + 1 jet
    • request without filter
    • request more events
    • use AF2
  • control file --> MadGraph --> Madspin --> Pythia8
  • We generated 50000 events after applying filter.
  • We might produce 150000 events without filter for new MC production
  • Still use MC15
  • What trigger is used?
  • The paper says
  • Comparing with 0 jet, the cross-section becomes factor of 10 lower for the sample with 1 jets
  • Monojet with pT > 80 GeV
  • Higgsino study request >= 1 jet with pT > 30 GeV
  • Prepare all background in SR
  • Make N-1 plot
  • Check cross-section for Higgsino
  • Add the NUHM2 motivation into the internal note
  • Can copy from SS/3L internal note and modify it.

[ytatus94]

• 20170629 NUHM2 short term plans:

1) Shape of the mll distribution

==> mll distribution has the same shape as Higgsino when MadGraph and N2->llN1 is used.

2) Prepare jobOptions

• Compressed+compressed production
• Decays with MadSpin vs MadGraph vs Pythia
• Different decays, All vs N2->llN1

==> Use MadSpin (faster) or MadGraph with N2->llN1 and C1->ffN1. Use cross-section x BR in event weighting.

==> No need to use N2->WC1, C1->ffN1 production, it has very small branching fraction ~ 0.0045.

3) Production N2C1p, N2C1m, N2N1, N2N2, C1C1

==> Use N2C1p, N2C1m, N2N1, C1C1.

==> No need to use N2N2 as the cross-section is very low.

==> No need to use C1N1 production although it has a large cross-section, because it has 1 lepton and will not pass the 2 lepton selection.

4) Interpretation

[ytatus94]

• 20170726
  • Judita provides new job option
  • I run all mass points + 4 decays using new job option.
  • multiplier = 100 (for 2L3), 20 (for MET50)
  • maxEvents=10k
  • Also use maxEvents=20k for m12=700 4 decays to test.
  • Judita will calculate how many events we need to request.
  • MC production --> ready --> Joey's Higgsino common framework --> root file --> my cutflow --> calculate the significance Z.
  • Use the TRUTH3 to calculate the significance before we have official MC sample.
  • Prepare HistFitter config and produce exclusion plot.

[ytatus94]

• 20170731
  • We can use the Table 8 Higgsino optimise SR to do the study.
  • We have to use the weighted filter efficiency.
  • For the acceptance, we have to use the weighted events
  • BR(C1->l vl n1) ~ 0.1 (Mike will give me the value they use)