| 3 |
|
We have developed two data-driven methods to |
| 4 |
|
estimate the background in the signal region. |
| 5 |
|
The first one exploits the fact that |
| 6 |
< |
\met and \met$/\sqrt{\rm SumJetPt}$ are nearly |
| 6 |
> |
SumJetPt and \met$/\sqrt{\rm SumJetPt}$ are nearly |
| 7 |
|
uncorrelated for the $t\bar{t}$ background |
| 8 |
|
(Section~\ref{sec:abcd}); the second one |
| 9 |
|
is based on the fact that in $t\bar{t}$ the |
| 21 |
|
\subsection{ABCD method} |
| 22 |
|
\label{sec:abcd} |
| 23 |
|
|
| 24 |
< |
We find that in $t\bar{t}$ events \met and |
| 25 |
< |
\met$/\sqrt{\rm SumJetPt}$ are nearly uncorrelated. |
| 26 |
< |
This is demonstrated in Figure~\ref{fig:uncor}. |
| 24 |
> |
We find that in $t\bar{t}$ events SumJetPt and |
| 25 |
> |
\met$/\sqrt{\rm SumJetPt}$ are nearly uncorrelated, |
| 26 |
> |
as demonstrated in Fig.~\ref{fig:uncor}. |
| 27 |
|
Thus, we can use an ABCD method in the \met$/\sqrt{\rm SumJetPt}$ vs |
| 28 |
|
sumJetPt plane to estimate the background in a data driven way. |
| 29 |
|
|
| 30 |
< |
\begin{figure}[tb] |
| 30 |
> |
%\begin{figure}[bht] |
| 31 |
> |
%\begin{center} |
| 32 |
> |
%\includegraphics[width=0.75\linewidth]{uncorrelated.pdf} |
| 33 |
> |
%\caption{\label{fig:uncor}\protect Distributions of SumJetPt |
| 34 |
> |
%in MC $t\bar{t}$ events for different intervals of |
| 35 |
> |
%MET$/\sqrt{\rm SumJetPt}$.} |
| 36 |
> |
%\end{center} |
| 37 |
> |
%\end{figure} |
| 38 |
> |
|
| 39 |
> |
\begin{figure}[bht] |
| 40 |
|
\begin{center} |
| 41 |
< |
\includegraphics[width=0.75\linewidth]{uncorrelated.pdf} |
| 41 |
> |
\includegraphics[width=0.75\linewidth]{uncor.png} |
| 42 |
|
\caption{\label{fig:uncor}\protect Distributions of SumJetPt |
| 43 |
|
in MC $t\bar{t}$ events for different intervals of |
| 44 |
< |
MET$/\sqrt{\rm SumJetPt}$.} |
| 44 |
> |
MET$/\sqrt{\rm SumJetPt}$. h1, h2, and h3 refer to the MET$/\sqrt{\rm SumJetPt}$ |
| 45 |
> |
intervals 4.5-6.5, 6.5-8.5 and $>$8.5, respectively. } |
| 46 |
|
\end{center} |
| 47 |
|
\end{figure} |
| 48 |
|
|
| 49 |
< |
\begin{figure}[bt] |
| 49 |
> |
\begin{figure}[tb] |
| 50 |
|
\begin{center} |
| 51 |
< |
\includegraphics[width=0.5\linewidth, angle=90]{abcdMC.pdf} |
| 52 |
< |
\caption{\label{fig:abcdMC}\protect Distributions of SumJetPt |
| 53 |
< |
vs. MET$/\sqrt{\rm SumJetPt}$ for SM Monte Carlo. Here we also |
| 54 |
< |
show our choice of ABCD regions.} |
| 51 |
> |
\includegraphics[width=0.75\linewidth]{ttdil_uncor_38X.png} |
| 52 |
> |
\caption{\label{fig:abcdMC}\protect Distributions of MET$/\sqrt{\rm SumJetPt}$ vs. |
| 53 |
> |
SumJetPt for SM Monte Carlo. Here we also show our choice of ABCD regions. The correlation coefficient |
| 54 |
> |
${\rm corr_{XY}}$ is computed for events falling in the ABCD regions.} |
| 55 |
|
\end{center} |
| 56 |
|
\end{figure} |
| 57 |
|
|
| 58 |
|
|
| 59 |
|
Our choice of ABCD regions is shown in Figure~\ref{fig:abcdMC}. |
| 60 |
|
The signal region is region D. The expected number of events |
| 61 |
< |
in the four regions for the SM Monte Carlo, as well as the BG |
| 62 |
< |
prediction AC/B are given in Table~\ref{tab:abcdMC} for an integrated |
| 63 |
< |
luminosity of 35 pb$^{-1}$. The ABCD method is accurate |
| 64 |
< |
to about 20\%. |
| 61 |
> |
in the four regions for the SM Monte Carlo, as well as the background |
| 62 |
> |
prediction A $\times$ C / B are given in Table~\ref{tab:abcdMC} for an integrated |
| 63 |
> |
luminosity of 34.0 pb$^{-1}$. In Table~\ref{tab:abcdsyst}, we test the stability of |
| 64 |
> |
observed/predicted with respect to variations in the ABCD boundaries. |
| 65 |
> |
Based on the results in Tables~\ref{tab:abcdMC} and~\ref{tab:abcdsyst}, we assess |
| 66 |
> |
a systematic uncertainty of 20\% on the prediction of the ABCD method. |
| 67 |
> |
|
| 68 |
> |
%As shown in Table~\ref{tab:abcdsyst}, we assess systematic uncertainties |
| 69 |
> |
%by varying the boundaries by an amount consistent with the hadronic energy scale uncertainty, |
| 70 |
> |
%which we take as $\pm$5\% for SumJetPt and $\pm$2.5\% for MET/$\sqrt{\rm SumJetPt}$, since the |
| 71 |
> |
%uncertainty on this quantity partially cancels due to the fact that it is a ratio of correlated |
| 72 |
> |
%quantities. Based on these studies we assess a correction factor $k_{ABCD} = 1.2 \pm 0.2$ to the |
| 73 |
> |
%predicted yield using the ABCD method. |
| 74 |
> |
|
| 75 |
> |
|
| 76 |
|
%{\color{red} Avi wants some statement about stability |
| 77 |
|
%wrt changes in regions. I am not sure that we have done it and |
| 78 |
|
%I am not sure it is necessary (Claudio).} |
| 79 |
|
|
| 80 |
< |
\begin{table}[htb] |
| 80 |
> |
\begin{table}[ht] |
| 81 |
|
\begin{center} |
| 82 |
|
\caption{\label{tab:abcdMC} Expected SM Monte Carlo yields for |
| 83 |
< |
35 pb$^{-1}$ in the ABCD regions, as well as the predicted yield in |
| 84 |
< |
the signal region given by A$\times$C/B. Here 'SM other' is the sum |
| 83 |
> |
34.0~pb$^{-1}$ in the ABCD regions, as well as the predicted yield in |
| 84 |
> |
the signal region given by A $\times$ C / B. Here `SM other' is the sum |
| 85 |
|
of non-dileptonic $t\bar{t}$ decays, $W^{\pm}$+jets, $W^+W^-$, |
| 86 |
|
$W^{\pm}Z^0$, $Z^0Z^0$ and single top.} |
| 87 |
< |
\begin{tabular}{l||c|c|c|c||c} |
| 87 |
> |
\begin{tabular}{lccccc} |
| 88 |
> |
%%%official json v3, 38X MC (D6T ttbar and DY) |
| 89 |
|
\hline |
| 90 |
< |
sample & A & B & C & D & A$\times$C/B \\ |
| 90 |
> |
sample & A & B & C & D & PRED \\ |
| 91 |
|
\hline |
| 92 |
< |
$t\bar{t}\rightarrow \ell^{+}\ell^{-}$ & 7.96 & 33.07 & 4.81 & 1.20 & 1.16 \\ |
| 93 |
< |
$Z^0$ + jets & 0.00 & 1.16 & 0.08 & 0.08 & 0.00 \\ |
| 94 |
< |
SM other & 0.65 & 2.31 & 0.17 & 0.14 & 0.05 \\ |
| 92 |
> |
$t\bar{t}\rightarrow \ell^{+}\ell^{-}$ & 8.44 $\pm$ 0.18 & 32.83 $\pm$ 0.35 & 4.78 $\pm$ 0.14 & 1.07 $\pm$ 0.06 & 1.23 $\pm$ 0.05 \\ |
| 93 |
> |
$Z^0 \rightarrow \ell^{+}\ell^{-}$ & 0.17 $\pm$ 0.08 & 1.18 $\pm$ 0.22 & 0.04 $\pm$ 0.04 & 0.12 $\pm$ 0.07 & 0.01 $\pm$ 0.01 \\ |
| 94 |
> |
SM other & 0.53 $\pm$ 0.03 & 2.26 $\pm$ 0.11 & 0.23 $\pm$ 0.03 & 0.07 $\pm$ 0.01 & 0.05 $\pm$ 0.01 \\ |
| 95 |
|
\hline |
| 96 |
< |
total SM MC & 8.61 & 36.54 & 5.05 & 1.41 & 1.19 \\ |
| 96 |
> |
total SM MC & 9.14 $\pm$ 0.20 & 36.26 $\pm$ 0.43 & 5.05 $\pm$ 0.14 & 1.27 $\pm$ 0.10 & 1.27 $\pm$ 0.05 \\ |
| 97 |
|
\hline |
| 98 |
|
\end{tabular} |
| 99 |
|
\end{center} |
| 100 |
|
\end{table} |
| 101 |
|
|
| 102 |
+ |
|
| 103 |
+ |
|
| 104 |
+ |
\begin{table}[ht] |
| 105 |
+ |
\begin{center} |
| 106 |
+ |
\caption{\label{tab:abcdsyst} |
| 107 |
+ |
Results of the systematic study of the ABCD method by varying the boundaries |
| 108 |
+ |
between the ABCD regions shown in Fig.~\ref{fig:abcdMC}. Here $x_1$ is the lower SumJetPt boundary and |
| 109 |
+ |
$x_2$ is the boundary separating regions A and B from C and D, their nominal values are 125 and 300~GeV, |
| 110 |
+ |
respectively. $y_1$ is the lower MET/$\sqrt{\rm SumJetPt}$ boundary and |
| 111 |
+ |
$y_2$ is the boundary separating regions B and C from A and D, their nominal values are 4.5 and 8.5~GeV$^{1/2}$, |
| 112 |
+ |
respectively.} |
| 113 |
+ |
\begin{tabular}{cccc|c} |
| 114 |
+ |
\hline |
| 115 |
+ |
$x_1$ & $x_2$ & $y_1$ & $y_2$ & Observed/Predicted \\ |
| 116 |
+ |
\hline |
| 117 |
+ |
|
| 118 |
+ |
nominal & nominal & nominal & nominal & $1.00 \pm 0.08$ \\ |
| 119 |
+ |
|
| 120 |
+ |
+5\% & +5\% & +2.5\% & +2.5\% & $1.08 \pm 0.11$ \\ |
| 121 |
+ |
|
| 122 |
+ |
+5\% & +5\% & nominal & nominal & $1.04 \pm 0.10$ \\ |
| 123 |
+ |
|
| 124 |
+ |
nominal & nominal & +2.5\% & +2.5\% & $1.03 \pm 0.09$ \\ |
| 125 |
+ |
|
| 126 |
+ |
nominal & +5\% & nominal & +2.5\% & $1.05 \pm 0.10$ \\ |
| 127 |
+ |
|
| 128 |
+ |
nominal & -5\% & nominal & -2.5\% & $0.95 \pm 0.07$ \\ |
| 129 |
+ |
|
| 130 |
+ |
-5\% & -5\% & +2.5\% & +2.5\% & $1.00 \pm 0.08$ \\ |
| 131 |
+ |
|
| 132 |
+ |
+5\% & +5\% & -2.5\% & -2.5\% & $0.98 \pm 0.09$ \\ |
| 133 |
+ |
\hline |
| 134 |
+ |
\end{tabular} |
| 135 |
+ |
\end{center} |
| 136 |
+ |
\end{table} |
| 137 |
+ |
|
| 138 |
+ |
|
| 139 |
+ |
\clearpage |
| 140 |
+ |
|
| 141 |
|
\subsection{Dilepton $P_T$ method} |
| 142 |
|
\label{sec:victory} |
| 143 |
|
This method is based on a suggestion by V. Pavlunin\cite{ref:victory}, |
| 154 |
|
to account for the fact that any dilepton selection must include a |
| 155 |
|
moderate \met cut in order to reduce Drell Yan backgrounds. This |
| 156 |
|
is discussed in Section 5.3 of Reference~\cite{ref:ourvictory}; for a \met |
| 157 |
< |
cut of 50 GeV, the rescaling factor is obtained from the data as |
| 157 |
> |
cut of 50 GeV, the rescaling factor is obtained from the MC as |
| 158 |
|
|
| 159 |
|
\newcommand{\ptll} {\ensuremath{P_T(\ell\ell)}} |
| 160 |
|
\begin{center} |
| 161 |
< |
$ K = \frac{\int_0^{\infty} {\cal N}(\ptll)~~d\ptll~}{\int_{50}^{\infty} {\cal N}(\ptll)~~d\ptll~}$ |
| 161 |
> |
$ K = \frac{\int_0^{\infty} {\cal N}(\ptll)~~d\ptll~}{\int_{50}^{\infty} {\cal N}(\ptll)~~d\ptll~} = 1.5$ |
| 162 |
|
\end{center} |
| 163 |
|
|
| 164 |
|
|
| 104 |
– |
Monte Carlo studies give values of $K$ that are typically between 1.5 and 1.6, |
| 105 |
– |
depending on selection details. |
| 165 |
|
%%%TO BE REPLACED |
| 166 |
|
%Given the integrated luminosity of the |
| 167 |
|
%present dataset, the determination of $K$ in data is severely statistics |
| 173 |
|
|
| 174 |
|
%\noindent {\color{red} For the 11 pb result we have used $K$ from data.} |
| 175 |
|
|
| 176 |
+ |
|
| 177 |
+ |
\begin{figure}[bht] |
| 178 |
+ |
\begin{center} |
| 179 |
+ |
\includegraphics[width=0.75\linewidth]{genvictory_Dec13.png} |
| 180 |
+ |
\caption{\label{fig:genvictory}\protect Distributions $P_T(\ell \ell)$ |
| 181 |
+ |
and $P_T(\nu \nu)$ (aka {\it genmet}) |
| 182 |
+ |
in $t\bar{t} \to$ dilepton Monte Carlo at the |
| 183 |
+ |
generator level. Events with $W \to \tau \to \ell$ are not included. |
| 184 |
+ |
No kinematical requirements have been made.} |
| 185 |
+ |
\end{center} |
| 186 |
+ |
\end{figure} |
| 187 |
+ |
|
| 188 |
+ |
|
| 189 |
|
There are several effects that spoil the correspondance between \met and |
| 190 |
|
$P_T(\ell\ell)$: |
| 191 |
|
\begin{itemize} |
| 192 |
|
\item $Ws$ in top events are polarized. Neutrinos are emitted preferentially |
| 193 |
< |
forward in the $W$ rest frame, thus the $P_T(\nu\nu)$ distribution is harder |
| 193 |
> |
parallel to the $W$ velocity while charged leptons are emitted prefertially |
| 194 |
> |
anti-parallel. Thus the $P_T(\nu\nu)$ distribution is harder |
| 195 |
|
than the $P_T(\ell\ell)$ distribution for top dilepton events. |
| 196 |
+ |
This turns out to be the dominant effect and it is illustrated in |
| 197 |
+ |
Figure~\ref{fig:genvictory}. |
| 198 |
|
\item The lepton selections results in $P_T$ and $\eta$ cuts on the individual |
| 199 |
|
leptons that have no simple correspondance to the neutrino requirements. |
| 200 |
|
\item Similarly, the \met$>$50 GeV cut introduces an asymmetry between leptons and |
| 201 |
|
neutrinos which is only partially compensated by the $K$ factor above. |
| 202 |
|
\item The \met resolution is much worse than the dilepton $P_T$ resolution. |
| 203 |
< |
When convoluted with a falling spectrum in the tails of \met, this result |
| 203 |
> |
When convoluted with a falling spectrum in the tails of \met, this results |
| 204 |
|
in a harder spectrum for \met than the original $P_T(\nu\nu)$. |
| 205 |
|
\item The \met response in CMS is not exactly 1. This causes a distortion |
| 206 |
|
in the \met distribution that is not present in the $P_T(\ell\ell)$ distribution. |
| 211 |
|
sources. These events can affect the background prediction. Particularly |
| 212 |
|
dangerous are high $P_T$ Drell Yan events that barely pass the \met$>$ 50 |
| 213 |
|
GeV selection. They will tend to push the data-driven background prediction up. |
| 214 |
+ |
Therefore we estimate the number of DY events entering the background prediction |
| 215 |
+ |
using the $R_{out/in}$ method as described in Sec.~\ref{sec:othBG}. |
| 216 |
|
\end{itemize} |
| 217 |
|
|
| 218 |
|
We have studied these effects in SM Monte Carlo, using a mixture of generator and |
| 219 |
|
reconstruction level studies, putting the various effects in one at a time. |
| 220 |
|
For each configuration, we apply the data-driven method and report as figure |
| 221 |
|
of merit the ratio of observed and predicted events in the signal region. |
| 222 |
< |
The results are summarized in Table~\ref{tab:victorybad}. |
| 222 |
> |
The figure of merit is calculated as follows |
| 223 |
> |
\begin{itemize} |
| 224 |
> |
\item We construct \met/$\sqrt{{\rm sumJetPt}}$ |
| 225 |
> |
and $P_T(\ell \ell)/\sqrt{{\rm sumJetPt}}$ (rescaled by the factor $K$ defined |
| 226 |
> |
above) distributions. |
| 227 |
> |
\item The distributions are constructed using either |
| 228 |
> |
GEN or RECO, and including or excluding various effects ({\em e.g.:} |
| 229 |
> |
$t \to W \to \tau \to \ell$). |
| 230 |
> |
\item In all cases the $N_{jets} \ge 2$ and |
| 231 |
> |
sumJetPt $>$ 300 GeV requirements are applied. |
| 232 |
> |
\item ``observed events'' is the integral of the \met/$\sqrt{{\rm sumJetPt}}$ distribution |
| 233 |
> |
above 8.5. |
| 234 |
> |
\item ``predicted events'' is the integral of the $P_T(\ell \ell)/\sqrt{{\rm sumJetPt}}$ distribution |
| 235 |
> |
above 8.5. |
| 236 |
> |
\end{itemize} |
| 237 |
> |
The results are summarized in Table~\ref{tab:victorybad}. Distributions corresponding to |
| 238 |
> |
lines 4 and 5 of Table~\ref{tab:victorybad} are shown in Figure~\ref{fig:victorybad}. |
| 239 |
|
|
| 240 |
|
\begin{table}[htb] |
| 241 |
|
\begin{center} |
| 242 |
< |
\caption{\label{tab:victorybad} Test of the data driven method in Monte Carlo |
| 243 |
< |
under different assumptions. See text for details.} |
| 242 |
> |
\caption{\label{tab:victorybad} |
| 243 |
> |
Test of the data driven method in Monte Carlo |
| 244 |
> |
under different assumptions, evaluated using Spring10 MC. See text for details.} |
| 245 |
|
\begin{tabular}{|l|c|c|c|c|c|c|c|c|} |
| 246 |
|
\hline |
| 247 |
|
& True $t\bar{t}$ dilepton & $t\to W\to\tau$& other SM & GEN or & Lepton $P_T$ & Z veto & \met $>$ 50& obs/pred \\ |
| 248 |
< |
& included & included & included & RECOSIM & and $\eta$ cuts & & & \\ \hline |
| 248 |
> |
& included & included & included & RECOSIM & and $\eta$ cuts & & & \\ \hline |
| 249 |
|
1&Y & N & N & GEN & N & N & N & 1.90 \\ |
| 250 |
|
2&Y & N & N & GEN & Y & N & N & 1.64 \\ |
| 251 |
|
3&Y & N & N & GEN & Y & Y & N & 1.59 \\ |
| 252 |
|
4&Y & N & N & GEN & Y & Y & Y & 1.55 \\ |
| 253 |
|
5&Y & N & N & RECOSIM & Y & Y & Y & 1.51 \\ |
| 254 |
|
6&Y & Y & N & RECOSIM & Y & Y & Y & 1.58 \\ |
| 255 |
< |
7&Y & Y & Y & RECOSIM & Y & Y & Y & 1.18 \\ |
| 255 |
> |
7&Y & Y & Y & RECOSIM & Y & Y & Y & 1.38 \\ |
| 256 |
|
\hline |
| 257 |
|
\end{tabular} |
| 258 |
|
\end{center} |
| 259 |
|
\end{table} |
| 260 |
|
|
| 261 |
|
|
| 262 |
+ |
\begin{figure}[bht] |
| 263 |
+ |
\begin{center} |
| 264 |
+ |
\includegraphics[width=0.48\linewidth]{genvictory_sqrtHt_Dec13.png} |
| 265 |
+ |
\includegraphics[width=0.48\linewidth]{victory_Dec13.png} |
| 266 |
+ |
\caption{\label{fig:victorybad}\protect Distributions |
| 267 |
+ |
of MET/$\sqrt{{\rm sumJetPt}}$ (black) and $P_T(\ell \ell)/\sqrt{{\rm sumJetPt}}$ |
| 268 |
+ |
(red) in $t\bar{t} \to$ dilepton Monte Carlo |
| 269 |
+ |
after lepton kinematical cuts, $N_{jets} \ge 2$, and |
| 270 |
+ |
sumJetPt $>$ 300 GeV. The left (right) plot is at the GEN (RECO) level |
| 271 |
+ |
and corresponds to line 4 (5) of Table~\ref{tab:victorybad}.} |
| 272 |
+ |
\end{center} |
| 273 |
+ |
\end{figure} |
| 274 |
+ |
|
| 275 |
+ |
|
| 276 |
+ |
|
| 277 |
+ |
\begin{table}[htb] |
| 278 |
+ |
\begin{center} |
| 279 |
+ |
\caption{\label{tab:victorysyst} |
| 280 |
+ |
Summary of variations in $K_C$ due to the MET scale and resolution uncertainty, and to backgrounds other than $t\bar{t} \to$~dilepton. |
| 281 |
+ |
In the first table, `up' and `down' refer to shifting the hadronic energy scale up and down by 5\%. In the second table, the quoted value |
| 282 |
+ |
refers to the amount of additional smearing of the MET, as discussed in the text. In the third table, the normalization of all backgrounds |
| 283 |
+ |
other than $t\bar{t} \to$~dilepton is varied. } |
| 284 |
+ |
\begin{tabular}{ lcccc } |
| 285 |
+ |
\hline |
| 286 |
+ |
MET scale & Predicted & Observed & Obs/pred \\ |
| 287 |
+ |
\hline |
| 288 |
+ |
nominal & 0.92 $ \pm $ 0.09 & 1.27 $ \pm $ 0.10 & 1.39 $ \pm $ 0.18 \\ |
| 289 |
+ |
up & 0.90 $ \pm $ 0.09 & 1.58 $ \pm $ 0.10 & 1.75 $ \pm $ 0.21 \\ |
| 290 |
+ |
down & 0.70 $ \pm $ 0.06 & 0.96 $ \pm $ 0.09 & 1.37 $ \pm $ 0.18 \\ |
| 291 |
+ |
\hline |
| 292 |
+ |
MET smearing & Predicted & Observed & Obs/pred \\ |
| 293 |
+ |
\hline |
| 294 |
+ |
nominal & 0.92 $ \pm $ 0.09 & 1.27 $ \pm $ 0.10 & 1.39 $ \pm $ 0.18 \\ |
| 295 |
+ |
10\% & 0.88 $ \pm $ 0.09 & 1.28 $ \pm $ 0.10 & 1.47 $ \pm $ 0.19 \\ |
| 296 |
+ |
20\% & 0.87 $ \pm $ 0.09 & 1.26 $ \pm $ 0.10 & 1.44 $ \pm $ 0.19 \\ |
| 297 |
+ |
30\% & 1.03 $ \pm $ 0.17 & 1.33 $ \pm $ 0.10 & 1.29 $ \pm $ 0.23 \\ |
| 298 |
+ |
40\% & 0.88 $ \pm $ 0.09 & 1.36 $ \pm $ 0.10 & 1.55 $ \pm $ 0.20 \\ |
| 299 |
+ |
50\% & 0.80 $ \pm $ 0.07 & 1.39 $ \pm $ 0.10 & 1.73 $ \pm $ 0.19 \\ |
| 300 |
+ |
\hline |
| 301 |
+ |
non-$t\bar{t} \to$~dilepton bkg & Predicted & Observed & Obs/pred \\ |
| 302 |
+ |
\hline |
| 303 |
+ |
ttdil only & 0.79 $ \pm $ 0.07 & 1.07 $ \pm $ 0.06 & 1.36 $ \pm $ 0.14 \\ |
| 304 |
+ |
nominal & 0.92 $ \pm $ 0.09 & 1.27 $ \pm $ 0.10 & 1.39 $ \pm $ 0.18 \\ |
| 305 |
+ |
double non-ttdil yield & 1.04 $ \pm $ 0.15 & 1.47 $ \pm $ 0.16 & 1.40 $ \pm $ 0.25 \\ |
| 306 |
+ |
\hline |
| 307 |
+ |
\end{tabular} |
| 308 |
+ |
\end{center} |
| 309 |
+ |
\end{table} |
| 310 |
+ |
|
| 311 |
|
The largest discrepancy between prediction and observation occurs on the first |
| 312 |
|
line of Table~\ref{tab:victorybad}, {\em i.e.}, at the generator level with no |
| 313 |
|
cuts. We have verified that this effect is due to the polarization of |
| 319 |
|
% by $\approx 4\%$\footnote{We find that observed/predicted changes by roughly 0.1 |
| 320 |
|
%for each 1.5\% change in \met response.}. |
| 321 |
|
Finally, contamination from non $t\bar{t}$ |
| 322 |
< |
events can have a significant impact on the BG prediction. The changes between |
| 323 |
< |
lines 6 and 7 of Table~\ref{tab:victorybad} is driven by 3 |
| 324 |
< |
Drell Yan events that pass the \met selection in Monte Carlo (thus the effect |
| 325 |
< |
is statistically not well quantified). |
| 322 |
> |
events can have a significant impact on the BG prediction. |
| 323 |
> |
%The changes between |
| 324 |
> |
%lines 6 and 7 of Table~\ref{tab:victorybad} is driven by 3 |
| 325 |
> |
%Drell Yan events that pass the \met selection in Monte Carlo (thus the effect |
| 326 |
> |
%is statistically not well quantified). |
| 327 |
|
|
| 328 |
|
An additional source of concern is that the CMS Madgraph $t\bar{t}$ MC does |
| 329 |
|
not include effects of spin correlations between the two top quarks. |
| 330 |
|
We have studied this effect at the generator level using Alpgen. We find |
| 331 |
< |
that the bias is at the few percent level. |
| 188 |
< |
|
| 189 |
< |
%%%TO BE REPLACED |
| 190 |
< |
%Based on the results of Table~\ref{tab:victorybad}, we conclude that the |
| 191 |
< |
%naive data driven background estimate based on $P_T{(\ell\ell)}$ needs to |
| 192 |
< |
%be corrected by a factor of {\color{red} $ K_{\rm{fudge}} =1.2 \pm 0.3$ |
| 193 |
< |
%(We still need to settle on thie exact value of this. |
| 194 |
< |
%For the 11 pb analysis it is taken as =1.)} . The quoted |
| 195 |
< |
%uncertainty is based on the stability of the Monte Carlo tests under |
| 196 |
< |
%variations of event selections, choices of \met algorithm, etc. |
| 197 |
< |
%For example, we find that observed/predicted changes by roughly 0.1 |
| 198 |
< |
%for each 1.5\% change in the average \met response. |
| 199 |
< |
|
| 200 |
< |
Based on the results of Table~\ref{tab:victorybad}, we conclude that the |
| 201 |
< |
naive data driven background estimate based on $P_T{(\ell\ell)}$ needs to |
| 202 |
< |
be corrected by a factor of $ K_C = X \pm Y$. |
| 203 |
< |
The value of this correction factor as well as the systematic uncertainty |
| 204 |
< |
will be assessed using 38X ttbar madgraph MC. In the following we use |
| 205 |
< |
$K_C = 1$ for simplicity. Based on previous MC studies we foresee a correction |
| 206 |
< |
factor of $K_C \approx 1.2 - 1.5$, and we will assess an uncertainty |
| 207 |
< |
based on the stability of the Monte Carlo tests under |
| 208 |
< |
variations of event selections, choices of \met algorithm, etc. |
| 209 |
< |
For example, we find that observed/predicted changes by roughly 0.1 |
| 210 |
< |
for each 1.5\% change in the average \met response. |
| 331 |
> |
that the bias is (at most) at the few percent level. |
| 332 |
|
|
| 333 |
+ |
Based on the results of Table~\ref{tab:victorysyst}, we conclude that the |
| 334 |
+ |
naive data-driven background estimate based on $P_T{(\ell\ell)}$ needs to |
| 335 |
+ |
be corrected by a factor of $ K_C = 1.4 \pm 0.2({\rm stat})$. |
| 336 |
+ |
|
| 337 |
+ |
The dominant sources of systematic uncertainty in $K_C$ are due to non-$t\bar{t} \to$~dilepton backgrounds, |
| 338 |
+ |
and the MET scale and resolution uncertainties, as summarized in Table~\ref{tab:victorysyst}. |
| 339 |
+ |
The impact of non-$t\bar{t}$-dilepton background is assessed |
| 340 |
+ |
by varying the yield of all backgrounds except for $t\bar{t} \to$~dilepton. |
| 341 |
+ |
The uncertainty is assessed as the larger of the differences between the nominal $K_C$ value and the values |
| 342 |
+ |
obtained using only $t\bar{t} \to$~dilepton MC and obtained by doubling the non $t\bar{t} \to$~dilepton component, |
| 343 |
+ |
giving an uncertainty of $0.03$. |
| 344 |
+ |
|
| 345 |
+ |
The uncertainty in $K_C$ due to the MET scale uncertainty is assessed by varying the hadronic energy scale using |
| 346 |
+ |
the same method as in~\cite{ref:top}, giving an uncertainty of 0.36. |
| 347 |
+ |
We also assess the impact of the MET resolution |
| 348 |
+ |
uncertainty on $K_C$ by applying a random smearing to the MET. For each event, we determine the expected MET resolution |
| 349 |
+ |
based on the sumJetPt, and smear the MET to simulate an increase in the resolution of 10\%, 20\%, 30\%, 40\% and 50\%. |
| 350 |
+ |
The results show that $K_C$ does not depend strongly on the MET resolution and we therefore do not assess any uncertainty. |
| 351 |
|
|
| 352 |
+ |
Incorporating all the statistical and systematic uncertainties we find $K_C = 1.4 \pm 0.4$. |
| 353 |
|
|
| 354 |
|
\subsection{Signal Contamination} |
| 355 |
|
\label{sec:sigcont} |
| 383 |
|
\caption{\label{tab:sigcont} Effects of signal contamination |
| 384 |
|
for the two data-driven background estimates. The three columns give |
| 385 |
|
the expected yield in the signal region and the background estimates |
| 386 |
< |
using the ABCD and $P_T(\ell \ell)$ methods. Results are normalized to 35~pb$^{-1}$.} |
| 386 |
> |
using the ABCD and $P_T(\ell \ell)$ methods. Results are normalized to 34.0~pb$^{-1}$.} |
| 387 |
|
\begin{tabular}{lccc} |
| 388 |
|
\hline |
| 389 |
|
& Yield & ABCD & $P_T(\ell \ell)$ \\ |
| 390 |
|
\hline |
| 391 |
< |
SM only & 1.41 & 1.19 & 0.96 \\ |
| 392 |
< |
SM + LM0 & 7.88 & 4.24 & 2.28 \\ |
| 393 |
< |
SM + LM1 & 3.98 & 1.53 & 1.44 \\ |
| 391 |
> |
SM only & 1.3 & 1.3 & 0.9 \\ |
| 392 |
> |
SM + LM0 & 9.9 & 6.1 & 2.4 \\ |
| 393 |
> |
SM + LM1 & 4.8 & 1.8 & 1.6 \\ |
| 394 |
> |
%SM only & 1.27 & 1.27 & 0.92 \\ |
| 395 |
> |
%SM + LM0 & 7.39 & 4.38 & 1.93 \\ |
| 396 |
> |
%SM + LM1 & 3.77 & 1.62 & 1.41 \\ |
| 397 |
|
\hline |
| 398 |
|
\end{tabular} |
| 399 |
|
\end{center} |
| 400 |
|
\end{table} |
| 401 |
|
|
| 259 |
– |
|
| 260 |
– |
|
| 261 |
– |
%\begin{table}[htb] |
| 262 |
– |
%\begin{center} |
| 263 |
– |
%\caption{\label{tab:sigcontABCD} Effects of signal contamination |
| 264 |
– |
%for the background predictions of the ABCD method including LM0 or |
| 265 |
– |
%LM1. Results |
| 266 |
– |
%are normalized to 30 pb$^{-1}$.} |
| 267 |
– |
%\begin{tabular}{|c|c||c|c||c|c|} |
| 268 |
– |
%\hline |
| 269 |
– |
%SM & BG Prediction & SM$+$LM0 & BG Prediction & SM$+$LM1 & BG Prediction \\ |
| 270 |
– |
%Background & SM Only & Contribution & Including LM0 & Contribution & Including LM1 \\ \hline |
| 271 |
– |
%1.2 & 1.0 & 6.8 & 3.7 & 3.4 & 1.3 \\ |
| 272 |
– |
%\hline |
| 273 |
– |
%\end{tabular} |
| 274 |
– |
%\end{center} |
| 275 |
– |
%\end{table} |
| 276 |
– |
|
| 277 |
– |
%\begin{table}[htb] |
| 278 |
– |
%\begin{center} |
| 279 |
– |
%\caption{\label{tab:sigcontPT} Effects of signal contamination |
| 280 |
– |
%for the background predictions of the $P_T(\ell\ell)$ method including LM0 or |
| 281 |
– |
%LM1. Results |
| 282 |
– |
%are normalized to 30 pb$^{-1}$.} |
| 283 |
– |
%\begin{tabular}{|c|c||c|c||c|c|} |
| 284 |
– |
%\hline |
| 285 |
– |
%SM & BG Prediction & SM$+$LM0 & BG Prediction & SM$+$LM1 & BG Prediction \\ |
| 286 |
– |
%Background & SM Only & Contribution & Including LM0 & Contribution & Including LM1 \\ \hline |
| 287 |
– |
%1.2 & 1.0 & 6.8 & 2.2 & 3.4 & 1.5 \\ |
| 288 |
– |
%\hline |
| 289 |
– |
%\end{tabular} |
| 290 |
– |
%\end{center} |
| 291 |
– |
%\end{table} |
| 292 |
– |
|
