Difference: CMSSusyAnalysisResults (1 vs. 25)

• New MC/Gen Info in 74X - PPD General - June 17, 2015

• Boosted top trigger - AK8PFJet360_TrimMass30 - ,March 26, 2015

• MC production: Run II MC Requests [All requests] [Main requests] --> Spreadsheets: [1], SUSY Slides: MC plans

• 15/04/15 - Razor meeting
• 15/03/29 - JHU elog #234 - Background estimation (NTop sideband, tau32 inversion)
• 15/03/16 - JHU elog #232 - Background Estimation (DPhi/NTop sideband - tau32/mass inversion)
• 15/03/09 - JHU elog #231 -

• • Has at least 2 AK8 jets with pt > 400, prunedMass > 140
  • Sideband cut: NTopHad < 2 (at least one tau32 cut is inverted)

• MC production: Run II MC Requests [All requests] [Main requests] --> Spreadsheets: [1] [2], SUSY Slides: MC plans

Twikis:

• SM cross sections
• MC production: Run II MC Requests --> Spreadsheet SUSY plans
• [[][]]
• [[][]]

Below test of clusure with slightly relaxed R cut:

• R cut is relaxed to > 0.35

given

• R cut is relaxed to > 0.35
• No lepton veto applied
• For TTBar and QCD, Sideband is in the DPhi>2.8 region instead of NTop<2

• Loose lepton veto

Signal selection cuts:

• 2 Hadronic tops
• Loose lepton veto

Introducing the method

• Define NTop<2 sideband, where the sideband is defined as 2 loose top-like jets:
  • Has at least 2 AK8 jets with pt > 400, prunedMass > 100
  • Sideband cut: NTopHad < 2
• Measure event yields in the R<0.4 sideband both for A) NTop<2 and C) NTop==2
• Fit NTop<2 sideband (A, B) with an exponential decay curve: exp(const + x * slope)
  • Extract fit parameters and their error: const, slope
• Scale function by the event yield ratio C/A to get Signal region prediction: D

Plots - Extrapolation of R shape using an exponential fit in the NTop<2 SideBand

TTbar	QCD	Comment
		R shape fit with exponential in the Ntop<2 sideband and NTop==2 signal bands. Range [0.15,1.0] for sideband, [0.15,0.4] for signal band.
W+Jets	Z+Jets --> NuNu	Comment
		R shape fit with exponential in the Ntop<2 sideband and NTop==2 signal bands. Range [0.15,1.0] for sideband, [0.15,0.4] for signal band, except for WJets for which low range is 0.2.
G + Jets	Single top (tW channel)	Comment
		R shape fit with exponential in the Ntop<2 sideband and NTop==2 signal bands. Range [0.15,1.0] for sideband, [0.15,0.4] for signal band.
DY + Jets	Comment
	R shape fit with exponential in the Ntop<2 sideband and NTop==2 signal bands. Range [0.15,1.0] for sideband, [0.15,0.4] for signal band.

Table

• Top Tagging at 13 TeV - JMAR Meeting, March 11, 2015

TTbar	QCD	W+Jets	Comment
			R shape fit with exponential in the DPhi (<2.8) Signal and (>2.8) side bands. Range [0.1,1.0] for sideband, [0.1,0.4] for signal band.
Z+Jets --> NuNu	G + Jets	Single top (tW channel)	Comment
			R shape fit with exponential in the DPhi (<2.8) Signal and (>2.8) side bands. Range [0.1,1.0] for sideband, [0.1,0.4] for signal band. For ZJetsToNuNu [0.15,0.4].

Resutls:

• Events only produce 2 hadronic top-tagged jets in high HT bins (eg.: 5-600 and above)
• Only O(10^3) events pass 2 hadronic top cut.
• DeltaPhi as a sideband may not work, because events with genuine missing ET seems to have correlations between DPhi and R, DPhi cut needs to be applied

Signal discrimination is good, but we could improve statistical error by applying looser cuts and fit shape in a sideband. This is also good to check the validity of the method.

--> Tried fit in the NTop sideband, where the sideband is defined as (2 loose top-like jets):

• Leading and sub-leading jet: pt > 400, prunedMass > 100
• Not yet applied, but: NTophad < 2 cut makes it a proper sideband

W+Jets (HT600)	Z+Jets --> NuNu (HT600)	Comment
		R shape distributions and fit with exponential in NTop (2 loose top-like jets) sideband (top line), NHadTop ==2, NHadTop ==2 && DPhi<2.8. For top 2 lines the slope of fits agree within error. For W+Jets also application of DPhi cut agrees, but not for ZJets which suggests that a DPhi cut could be applied in the sideband too.

Recipe to plot by hand:

root -l
TChain c("c");
c.Add("/data/gridout/jkarancs/SusyAnalysis/B2G/TTreeNtuple/Feb28_edm_Feb20/WJetsToLNu_HT-600toInf/*.root/B2GTTreeMaker/B2GTree");
c.Draw("evt_R","jetAK8_Pt[0]>400&&jetAK8_Pt[1]>400&&jetAK8_prunedMass[0]>100&&jetAK8_prunedMass[1]>100&&evt_NTopHad<2");
c.Draw("evt_R","evt_NTopHad==2","SAME")
c.Draw("evt_R","evt_NTopHad==2&&evt_TTHadDPhi<2.8","SAME")

• R > 0.4
• DPhi < 2.8

Preliminary results - Extrapolation of R shape using an exponential fit in the R<0.4 SideBand

• 2 Hadronic tops
• Lepton veto

Best cuts using Smin maximisation script - New

• R > 0.41

• HTall > 1100

Using cuts by hand:

• R > 0.4
• DPhi < 2.8
• HTall > 1500

• Background Estimation Methods (ABCD - slide 8)

• Viktor veszpremi - PhD (Wigner, Budapest, HU)
• Petar Maksimovic - Prof. PhD (Johns Hopkins University, Baltimore, US)

• Physics performance and potentials of boosted topologies - CMS Physics week, February 24, 2014

Links	Date	Author	Title	Comment
arXiv:physics/0308063	2003	G. Punzi	Sensitivity of searches for new signals and its optimization
arXiv:1011.2268	2010	J. Thaler, K. V. Tilburg	Identifying Boosted Objects with N-subjettiness	tau variables (N-sj.), Mistag rates, tagging efficiencies
arXiv:1006.2833	2010	T. Plehn et. al	Stop Reconstruction with Tagged Tops
CMS-PAS-B2G-12-006 (Twiki)	2013	CMS	Search for Anomalous Top Quark Pair Production in the Boosted All-Hadronic Final State	Mistag rate of top tagging algo, Multijet bkg estimation
CMS-PAS-B2G-12-006 (Twiki)	2013	CMS	Search for ttbar resonances in semileptonic final state	2D cut for semi-leptonic top
CMS-PAS-JME-13-007	2014 february	CMS	Boosted Top Jet Tagging at CMS
SUS-14-007 (Approval)	2015	Razor group	Search for SUSY with boosted W's using the razor variables

Presentations:

• Update on Differential tt Cross Section with Boosted Tops - Top Cross Section Meeting, July 24, 2014
• Jet Substructure Hands-on Advanced Tutorial Session - LPC, June 18-19, 2014
• QCD-Multijet Background Modelling Using The ABCD Method - ttH meeting, February 28, 2013

• arXiv:physics/0308063 (2003) - G. Punzi - Sensitivity of searches for new signals and its optimization
• CMS-PAS-JME-13-007 (2014 february) - CMS - Boosted Top Jet Tagging at CMS
• arXiv:1011.2268 (2010) - J. Thaler, K. V. Tilburg - Identifying Boosted Objects with N-subjettiness

Object/Event selections

Object definitions

MiniAOD objects

Jets

• pT > 100 GeV /c
• | eta | < 4

Electrons

• pT > 35 GeV /c
• | eta | < 2.5

Muons

• pT > 45 GeV /c
• | eta | < 2.1
• tight ID

MET

Event Selection

• pT > 400 GeV /c
• tau3 / tau2 < 0.75
• Pruned Mass > 140 Gev/c^2

• R (Razor variable) - Recipe
  • Calculated for jets with | eta | < 3
• H_T,all = Sum ( AK8Jet pT ) + Sum ( Lepton pT ) + MET
• DPhi = | DeltaPhi ( top_had1, top_had2 ) |

• R > 0.35
• DPhi < 2.8
• HTall > 1500

Expected number of events @ 10 fb^-1 @ Mgluino = 1.3 TeV : 460.525

• R > 0.32
• DPhi < 2.8
• HTall > 1450

Cut flow chart

Basic cuts by hand - presented on 25 February 2015

1. R > 0.35
2. DPhi < 2.8
3. HTall > 1500

Best cuts by Smin maximisation - New

1. R > 0.32
2. DPhi < 2.8
3. HTall > 1450