claudioc/OSNote2010/eventsel.tex

\section{Event Preselection}
\label{sec:eventSel}
The purpose of the preselection is to define a data sample rich 
in $t\bar{t} \to$ dileptons.  We compare the kinematical 
properties of this sample with expectations from $t\bar{t}$ 
Monte Carlo.

The preselection is based on the 
$t\bar{t}$ analysis~\cite{ref:top}.  
We select events with two opposite sign isolated
leptons ($ee$, $e\mu$, or $\mu\mu$); one of the leptons must 
have $P_T > 20$ GeV,
the other one must have $P_T > 10$ GeV. Events consistent with $Z$ are rejected.
In case of events with 
more than two such leptons, we select the pair that maximizes the scalar 
sum of lepton $P_T$'s.
There must be two JPT
jets of $P_T > 30$ GeV and $|\eta| < 2.5$; the scalar sum of the 
$P_T$ of all such jets must exceed 100 GeV; jets must pass
{\tt caloJetId} and be separated by $\Delta R >$ 0.4 from any 
lepton passing the selection.
Finally $\met > 50$ GeV (we use tcMet). More details are given in the subsections below.

\subsection{Event Cleanup}
\label{sec:cleanup}
\begin{itemize}
\item Scraping cut: if there are $\geq$ 10 tracks, require at
least 25\% of them to be high purity.
\item Require at least one good vertex:
\begin{itemize}
\item not fake
\item ndof $>$ 4
\item $|\rho| < 2$ cm
\item $|z| < 24$ cm.  
\end{itemize}
\end{itemize}


\subsection{Muon Selection}
\label{sec:muon}

Muon candidates are RECO muon objects passing the following
requirements:
\begin{itemize}

\item $|\eta| < 2.4$.

\item Global Muon and Tracker Muon.

\item $\chi^2$/ndof of global fit $<$ 10.

\item At least 11 hits in the tracker fit.

\item Transverse impact parameter with respect to the beamspot $<$ 200 $\mu$m.

\item $Iso \equiv $ $E_T^{\rm iso}$/Max(20 GeV, $P_T$) $<$ 0.15.  
$E_T^{\rm iso}$
is defined as the sum of transverse energy/momentum deposits in ecal,
hcal, and tracker, in a cone of 0.3.

\item At least one of the hits from the 
standalone muon must be used in the global fit.

\item Require tracker $\Delta P_T/P_T < 0.1$. This cut was not in the original top analysis.
It is motivated by the observation of
poorly measured muons in data with large
relative $P_T$ uncertainty, giving significant contributions to the \met.
%{\color{red} This is not applied to the 11 pb iteration.}


\end{itemize}


\subsection{Electron Selection}
\label{sec:electron}

Electron candidates are RECO GSF electrons passing the following
requirements:

\begin{itemize}

% \item $P_T > 10$ GeV.  (The $t\bar{t}$ analysis uses 20 GeV but for
% completeness we calculate FR down to 10 GeV).

\item $|\eta| < 2.5$.

\item SuperCluster $E_T > 10$ GeV.  

\item The electron must be ecal seeded.

\item VBTF90 identification\cite{ref:vbtf}.

\item Transverse impact parameter with respect to the beamspot $<$ 400 $\mu$m.

\item $Iso \equiv $ $E_T^{\rm iso}$/Max(20 GeV, $P_T$) $<$ 0.15.  
$E_T^{\rm iso}$
is defined as the sum of transverse energy/momentum deposits in ecal,
hcal, and tracker, in a 
cone of 0.3.  A 1 GeV pedestal is subtracted from the ecal energy 
deposition in the EB, however the ecal energy is never allowed to 
go negative.

\item Electrons with a tracker or global muon within $\Delta R$ of 
0.1 are vetoed.

\item The number of missing expected inner hits must be less than 
two\cite{ref:conv}.

\item Conversion removal via partner track finding: any electron
where an additional GeneralTrack is found with $Dist < 0.02$ cm 
and $\Delta \cot \theta < 0.02$ is vetoed\cite{ref:conv}.

\item Cleaning for ECAL spike (aka Swiss-Cross cleaning) has been applied
at the reconstruction level (CMSSW 38x).

\end{itemize}

\subsection{Invariant mass requirement}
\label{sec:zveto}

We remove $e^+e^-$ and $\mu^+ \mu^-$ events with invariant 
mass between 76 and 106 GeV.  We also remove events
with invariant mass $<$ 10 GeV, since this kinematical region is 
not well reprodced in CMS Monte Carlos.

In addition, we remove $Z \to \mu\mu\gamma$
candidates with the $\gamma$ collinear with one of the muons.  This is
done as follows:
if the ecal energy associated with one of the muons is greater than 6 GeV,
we add this energy to the momentum of the initial muon, and we recompute
the $\mu\mu$ mass.  If this mass is between 76 and 106 GeV, the event is rejected.


\subsection{Trigger Selection}
\label{sec:trigSel}

Because most of the triggers implemented in the 2nd half of the
2010 run were not implemented in the Monte Carlo, 
we do not make any requirements on HLT bits in the Monte Carlo.
Instead, as discussed in 
Section~\ref{sec:trgEff}, a trigger efficiency weight is applied
to each event, based on the trigger efficiencies measured on data.
Trigger efficiency weights are very close to 1.

%For data, we require the logical OR of all (or most?) unprescaled
%single and double lepton triggers that were deployed during the 2010
%run.  These are:
%{\color{red} Here we need to list the triggers, somehow.}

For data, we use a cocktail of unprescaled single 
and double lepton triggers. An event
in the $ee$ final state is required to pass at least 1 
single- or double-electron trigger, a
$\mu\mu$ event is required to pass at least 1 single 
or double-muon trigger, while an $e\mu$ event
is required to pass at least 1 single-muon, single-electron, 
or $e-\mu$ cross trigger. 
% We currently
% do not require MC events to pass any triggers.


\begin{itemize}
\item single-muon triggers
  \begin{itemize}
  \item \verb=HLT_Mu5=
  \item \verb=HLT_Mu7=       
  \item \verb=HLT_Mu9=        
  \item \verb=HLT_Mu11=      
  \item \verb=HLT_Mu13_v1=    
  \item \verb=HLT_Mu15_v1=    
  \item \verb=HLT_Mu17_v1=    
  \item \verb=HLT_Mu19_v1=    
  \end{itemize}
\item double-muon triggers
  \begin{itemize}
  \item \verb=HLT_DoubleMu3=
  \item \verb=HLT_DoubleMu3_v2=
  \item \verb=HLT_DoubleMu5_v1=
  \end{itemize}
\item single-electron triggers
  \begin{itemize}
  \item \verb=HLT_Ele10_SW_EleId_L1R=
  \item \verb=HLT_Ele10_LW_EleId_L1R=
  \item \verb=HLT_Ele10_LW_L1R=
  \item \verb=HLT_Ele10_SW_L1R=
  \item \verb=HLT_Ele15_SW_CaloEleId_L1R=
  \item \verb=HLT_Ele15_SW_EleId_L1R=
  \item \verb=HLT_Ele15_SW_L1R=
  \item \verb=HLT_Ele15_LW_L1R=
  \item \verb=HLT_Ele17_SW_TightEleId_L1R=
  \item \verb=HLT_Ele17_SW_TighterEleId_L1R_v1=
  \item \verb=HLT_Ele17_SW_CaloEleId_L1R=
  \item \verb=HLT_Ele17_SW_EleId_L1R=
  \item \verb=HLT_Ele17_SW_LooseEleId_L1R=
  \item \verb=HLT_Ele17_SW_TighterEleIdIsol_L1R_v2=
  \item \verb=HLT_Ele20_SW_L1R=
  \item \verb=HLT_Ele22_SW_TighterEleId_L1R_v2=
  \item \verb=HLT_Ele32_SW_TightCaloEleIdTrack_L1R_v1=
  \item \verb=HLT_Ele32_SW_TighterEleId_L1R_v2=
  \item \verb=HLT_Ele27_SW_TightCaloEleIdTrack_L1R_v1=
  \item \verb=HLT_Ele22_SW_TighterCaloIdIsol_L1R_v2=
  \item \verb=HLT_Ele22_SW_TighterEleId_L1R_v3=
  \item \verb=HLT_Ele22_SW_TighterCaloIdIsol_L1R_v2=
  \end{itemize}
\item double-electron triggers
  \begin{itemize}
  \item \verb=HLT_DoubleEle15_SW_L1R_v1=                
  \item \verb=HLT_DoubleEle17_SW_L1R_v1=  
  \item \verb=HLT_Ele17_SW_TightCaloEleId_Ele8HE_L1R_v1=
  \item \verb=HLT_Ele17_SW_TightCaloEleId_SC8HE_L1R_v1=
  \item \verb=HLT_DoubleEle10_SW_L1R=
  \item \verb=HLT_DoubleEle5_SW_L1R=
  \end{itemize}
\item e-$\mu$ cross triggers
  \begin{itemize}
  \item \verb=HLT_Mu5_Ele5_v1=
  \item \verb=HLT_Mu5_Ele9_v1=
  \item \verb=HLT_Mu11_Ele8_v1=
  \item \verb=HLT_Mu8_Ele8_v1=
  \item \verb=HLT_Mu5_Ele13_v2=
  \item \verb=HLT_Mu5_Ele17_v1=
  \end{itemize}
\end{itemize}
Revision:	1.12
Committed:	Thu Nov 11 16:22:13 2010 UTC (14 years, 6 months ago) by claudioc
Content type:	application/x-tex
Branch:	MAIN
Changes since 1.11:	+13 -7 lines
Log Message:	triv
#	Content
1	\section{Event Preselection}
2	\label{sec:eventSel}
3	The purpose of the preselection is to define a data sample rich
4	in $t\bar{t} \to$ dileptons. We compare the kinematical
5	properties of this sample with expectations from $t\bar{t}$
6	Monte Carlo.
7
8	The preselection is based on the
9	$t\bar{t}$ analysis~\cite{ref:top}.
10	We select events with two opposite sign isolated
11	leptons ($ee$, $e\mu$, or $\mu\mu$); one of the leptons must
12	have $P_T > 20$ GeV,
13	the other one must have $P_T > 10$ GeV. Events consistent with $Z$ are rejected.
14	In case of events with
15	more than two such leptons, we select the pair that maximizes the scalar
16	sum of lepton $P_T$'s.
17	There must be two JPT
18	jets of $P_T > 30$ GeV and $\|\eta\| < 2.5$; the scalar sum of the
19	$P_T$ of all such jets must exceed 100 GeV; jets must pass
20	{\tt caloJetId} and be separated by $\Delta R >$ 0.4 from any
21	lepton passing the selection.
22	Finally $\met > 50$ GeV (we use tcMet). More details are given in the subsections below.
23
24	\subsection{Event Cleanup}
25	\label{sec:cleanup}
26	\begin{itemize}
27	\item Scraping cut: if there are $\geq$ 10 tracks, require at
28	least 25\% of them to be high purity.
29	\item Require at least one good vertex:
30	\begin{itemize}
31	\item not fake
32	\item ndof $>$ 4
33	\item $\|\rho\| < 2$ cm
34	\item $\|z\| < 24$ cm.
35	\end{itemize}
36	\end{itemize}
37
38
39	\subsection{Muon Selection}
40	\label{sec:muon}
41
42	Muon candidates are RECO muon objects passing the following
43	requirements:
44	\begin{itemize}
45
46	\item $\|\eta\| < 2.4$.
47
48	\item Global Muon and Tracker Muon.
49
50	\item $\chi^2$/ndof of global fit $<$ 10.
51
52	\item At least 11 hits in the tracker fit.
53
54	\item Transverse impact parameter with respect to the beamspot $<$ 200 $\mu$m.
55
56	\item $Iso \equiv $ $E_T^{\rm iso}$/Max(20 GeV, $P_T$) $<$ 0.15.
57	$E_T^{\rm iso}$
58	is defined as the sum of transverse energy/momentum deposits in ecal,
59	hcal, and tracker, in a cone of 0.3.
60
61	\item At least one of the hits from the
62	standalone muon must be used in the global fit.
63
64	\item Require tracker $\Delta P_T/P_T < 0.1$. This cut was not in the original top analysis.
65	It is motivated by the observation of
66	poorly measured muons in data with large
67	relative $P_T$ uncertainty, giving significant contributions to the \met.
68	%{\color{red} This is not applied to the 11 pb iteration.}
69
70
71	\end{itemize}
72
73
74
75	\subsection{Electron Selection}
76	\label{sec:electron}
77
78	Electron candidates are RECO GSF electrons passing the following
79	requirements:
80
81	\begin{itemize}
82
83	% \item $P_T > 10$ GeV. (The $t\bar{t}$ analysis uses 20 GeV but for
84	% completeness we calculate FR down to 10 GeV).
85
86	\item $\|\eta\| < 2.5$.
87
88	\item SuperCluster $E_T > 10$ GeV.
89
90	\item The electron must be ecal seeded.
91
92	\item VBTF90 identification\cite{ref:vbtf}.
93
94	\item Transverse impact parameter with respect to the beamspot $<$ 400 $\mu$m.
95
96	\item $Iso \equiv $ $E_T^{\rm iso}$/Max(20 GeV, $P_T$) $<$ 0.15.
97	$E_T^{\rm iso}$
98	is defined as the sum of transverse energy/momentum deposits in ecal,
99	hcal, and tracker, in a
100	cone of 0.3. A 1 GeV pedestal is subtracted from the ecal energy
101	deposition in the EB, however the ecal energy is never allowed to
102	go negative.
103
104	\item Electrons with a tracker or global muon within $\Delta R$ of
105	0.1 are vetoed.
106
107	\item The number of missing expected inner hits must be less than
108	two\cite{ref:conv}.
109
110	\item Conversion removal via partner track finding: any electron
111	where an additional GeneralTrack is found with $Dist < 0.02$ cm
112	and $\Delta \cot \theta < 0.02$ is vetoed\cite{ref:conv}.
113
114	\item Cleaning for ECAL spike (aka Swiss-Cross cleaning) has been applied
115	at the reconstruction level (CMSSW 38x).
116
117	\end{itemize}
118
119	\subsection{Invariant mass requirement}
120	\label{sec:zveto}
121
122	We remove $e^+e^-$ and $\mu^+ \mu^-$ events with invariant
123	mass between 76 and 106 GeV. We also remove events
124	with invariant mass $<$ 10 GeV, since this kinematical region is
125	not well reprodced in CMS Monte Carlos.
126
127	In addition, we remove $Z \to \mu\mu\gamma$
128	candidates with the $\gamma$ collinear with one of the muons. This is
129	done as follows:
130	if the ecal energy associated with one of the muons is greater than 6 GeV,
131	we add this energy to the momentum of the initial muon, and we recompute
132	the $\mu\mu$ mass. If this mass is between 76 and 106 GeV, the event is rejected.
133
134
135	\subsection{Trigger Selection}
136	\label{sec:trigSel}
137
138	Because most of the triggers implemented in the 2nd half of the
139	2010 run were not implemented in the Monte Carlo,
140	we do not make any requirements on HLT bits in the Monte Carlo.
141	Instead, as discussed in
142	Section~\ref{sec:trgEff}, a trigger efficiency weight is applied
143	to each event, based on the trigger efficiencies measured on data.
144	Trigger efficiency weights are very close to 1.
145
146	%For data, we require the logical OR of all (or most?) unprescaled
147	%single and double lepton triggers that were deployed during the 2010
148	%run. These are:
149	%{\color{red} Here we need to list the triggers, somehow.}
150
151	For data, we use a cocktail of unprescaled single
152	and double lepton triggers. An event
153	in the $ee$ final state is required to pass at least 1
154	single- or double-electron trigger, a
155	$\mu\mu$ event is required to pass at least 1 single
156	or double-muon trigger, while an $e\mu$ event
157	is required to pass at least 1 single-muon, single-electron,
158	or $e-\mu$ cross trigger.
159	% We currently
160	% do not require MC events to pass any triggers.
161
162
163
164
165
166
167
168
169
170	\begin{itemize}
171	\item single-muon triggers
172	\begin{itemize}
173	\item \verb=HLT_Mu5=
174	\item \verb=HLT_Mu7=
175	\item \verb=HLT_Mu9=
176	\item \verb=HLT_Mu11=
177	\item \verb=HLT_Mu13_v1=
178	\item \verb=HLT_Mu15_v1=
179	\item \verb=HLT_Mu17_v1=
180	\item \verb=HLT_Mu19_v1=
181	\end{itemize}
182	\item double-muon triggers
183	\begin{itemize}
184	\item \verb=HLT_DoubleMu3=
185	\item \verb=HLT_DoubleMu3_v2=
186	\item \verb=HLT_DoubleMu5_v1=
187	\end{itemize}
188	\item single-electron triggers
189	\begin{itemize}
190	\item \verb=HLT_Ele10_SW_EleId_L1R=
191	\item \verb=HLT_Ele10_LW_EleId_L1R=
192	\item \verb=HLT_Ele10_LW_L1R=
193	\item \verb=HLT_Ele10_SW_L1R=
194	\item \verb=HLT_Ele15_SW_CaloEleId_L1R=
195	\item \verb=HLT_Ele15_SW_EleId_L1R=
196	\item \verb=HLT_Ele15_SW_L1R=
197	\item \verb=HLT_Ele15_LW_L1R=
198	\item \verb=HLT_Ele17_SW_TightEleId_L1R=
199	\item \verb=HLT_Ele17_SW_TighterEleId_L1R_v1=
200	\item \verb=HLT_Ele17_SW_CaloEleId_L1R=
201	\item \verb=HLT_Ele17_SW_EleId_L1R=
202	\item \verb=HLT_Ele17_SW_LooseEleId_L1R=
203	\item \verb=HLT_Ele17_SW_TighterEleIdIsol_L1R_v2=
204	\item \verb=HLT_Ele20_SW_L1R=
205	\item \verb=HLT_Ele22_SW_TighterEleId_L1R_v2=
206	\item \verb=HLT_Ele32_SW_TightCaloEleIdTrack_L1R_v1=
207	\item \verb=HLT_Ele32_SW_TighterEleId_L1R_v2=
208	\item \verb=HLT_Ele27_SW_TightCaloEleIdTrack_L1R_v1=
209	\item \verb=HLT_Ele22_SW_TighterCaloIdIsol_L1R_v2=
210	\item \verb=HLT_Ele22_SW_TighterEleId_L1R_v3=
211	\item \verb=HLT_Ele22_SW_TighterCaloIdIsol_L1R_v2=
212	\end{itemize}
213	\item double-electron triggers
214	\begin{itemize}
215	\item \verb=HLT_DoubleEle15_SW_L1R_v1=
216	\item \verb=HLT_DoubleEle17_SW_L1R_v1=
217	\item \verb=HLT_Ele17_SW_TightCaloEleId_Ele8HE_L1R_v1=
218	\item \verb=HLT_Ele17_SW_TightCaloEleId_SC8HE_L1R_v1=
219	\item \verb=HLT_DoubleEle10_SW_L1R=
220	\item \verb=HLT_DoubleEle5_SW_L1R=
221	\end{itemize}
222	\item e-$\mu$ cross triggers
223	\begin{itemize}
224	\item \verb=HLT_Mu5_Ele5_v1=
225	\item \verb=HLT_Mu5_Ele9_v1=
226	\item \verb=HLT_Mu11_Ele8_v1=
227	\item \verb=HLT_Mu8_Ele8_v1=
228	\item \verb=HLT_Mu5_Ele13_v2=
229	\item \verb=HLT_Mu5_Ele17_v1=
230	\end{itemize}
231	\end{itemize}