Upper limits on $\sigma\times B$ as a function of $m_{e^*}$ in the $e\nu J$ channel. The $\pm 1(2)\sigma$ uncertainty bands around the expected limit represent all sources of systematic and statistical uncertainties.

Lower limits on $\Lambda$ as a function of $m_{e^*}$ for the $eejj$ channel. The $\pm\ 1(2)\sigma$ uncertainty bands around the expected limit represent all sources of systematic and statistical uncertainties. The limits for $m_{e^*}\gt4$ TeV are the result of extrapolation.

Lower limits on $\Lambda$ as a function of $m_{e^*}$ for the $e\nu J$ channel. The $\pm\ 1(2)\sigma$ uncertainty bands around the expected limit represent all sources of systematic and statistical uncertainties.

Lower limits on $\Lambda$ as a function of $m_{e^*}$ combined for the $eejj$ and the $e\nu J$ channels. The $\pm\ 1(2)\sigma$ uncertainty bands around the expected limit represent all sources of systematic and statistical uncertainties. The limits for $m_{e^*}\gt4$ TeV are the result of extrapolation.

Lower limits at 95$\%$ CL on $\Lambda$ for the dimuon channel for different signal chiralities in the (constructive/destructive interference) SRs of the analysis.

Lower limits at 95$\%$ CL on $\Lambda$ for the dilepton combination for different signal chiralities in the (constructive/destructive interference) SRs of the analysis.

The dielectron data events passing the full selection.

The dimuon data events passing the full selection.

The absolute differential cross-section measured in the fiducial phase-space as a function of the minimum $\Delta R$ between the photon and the leptons in the electron-muon channel. The uncertainty is decomposed into four components which are the signal modelling uncertainty, the background modelling uncertainty, the experimental uncertainty, and the data statistical uncertainty.

The absolute differential cross-section measured in the fiducial phase-space as a function of the $\Delta\phi$ between the two leptons in the electron-muon channel. The uncertainty is decomposed into four components which are the signal modelling uncertainty, the background modelling uncertainty, the experimental uncertainty, and the data statistical uncertainty.

The absolute differential cross-section measured in the fiducial phase-space as a function of the $|\Delta\eta|$ between the two leptons in the electron-muon channel. The uncertainty is decomposed into four components which are the signal modelling uncertainty, the background modelling uncertainty, the experimental uncertainty, and the data statistical uncertainty.

The normalised differential cross-section measured in the fiducial phase-space as a function of the photon pT in the electron-muon channel. The uncertainty is decomposed into four components which are the signal modelling uncertainty, the background modelling uncertainty, the experimental uncertainty, and the data statistical uncertainty.

The normalised differential cross-section measured in the fiducial phase-space as a function of the photon $|\eta|$ in the electron-muon channel. The uncertainty is decomposed into four components which are the signal modelling uncertainty, the background modelling uncertainty, the experimental uncertainty, and the data statistical uncertainty.

The normalised differential cross-section measured in the fiducial phase-space as a function of the minimum $\Delta R$ between the photon and the leptons in the electron-muon channel. The uncertainty is decomposed into four components which are the signal modelling uncertainty, the background modelling uncertainty, the experimental uncertainty, and the data statistical uncertainty.

The normalised differential cross-section measured in the fiducial phase-space as a function of the $\Delta\phi$ between the two leptons in the electron-muon channel. The uncertainty is decomposed into four components which are the signal modelling uncertainty, the background modelling uncertainty, the experimental uncertainty, and the data statistical uncertainty.

The normalised differential cross-section measured in the fiducial phase-space as a function of the $|\Delta\eta|$ between the two leptons in the electron-muon channel. The uncertainty is decomposed into four components which are the signal modelling uncertainty, the background modelling uncertainty, the experimental uncertainty, and the data statistical uncertainty.

The total correlation matrix of the absolute differential cross-section measured in the fiducial phase-space as a function of the photon pT in the electron-muon channel. The individual systematic uncertainties are symmetrized before deriving the correlation matrix.

The total correlation matrix of the absolute differential cross-section measured in the fiducial phase-space as a function of the photon $|\eta|$ in the electron-muon channel. The individual systematic uncertainties are symmetrized before deriving the correlation matrix.

The total correlation matrix of the absolute differential cross-section measured in the fiducial phase-space as a function of the minimum $\Delta R$ between the photon and the leptons in the electron-muon channel. The individual systematic uncertainties are symmetrized before deriving the correlation matrix.

The total correlation matrix of the absolute differential cross-section measured in the fiducial phase-space as a function of the $\Delta\phi$ between the two leptons in the electron-muon channel. The individual systematic uncertainties are symmetrized before deriving the correlation matrix.

The total correlation matrix of the absolute differential cross-section measured in the fiducial phase-space as a function of the $|\Delta\eta|$ between the two leptons in the electron-muon channel. The individual systematic uncertainties are symmetrized before deriving the correlation matrix.

The total correlation matrix of the normalised differential cross-section measured in the fiducial phase-space as a function of the photon pT in the electron-muon channel. The individual systematic uncertainties are symmetrized before deriving the correlation matrix.

The total correlation matrix of the normalised differential cross-section measured in the fiducial phase-space as a function of the photon $|\eta|$ in the electron-muon channel. The individual systematic uncertainties are symmetrized before deriving the correlation matrix.

The total correlation matrix of the normalised differential cross-section measured in the fiducial phase-space as a function of the minimum $\Delta R$ between the photon and the leptons in the electron-muon channel. The individual systematic uncertainties are symmetrized before deriving the correlation matrix.

The total correlation matrix of the normalised differential cross-section measured in the fiducial phase-space as a function of the $\Delta\phi$ between the two leptons in the electron-muon channel. The individual systematic uncertainties are symmetrized before deriving the correlation matrix.

The total correlation matrix of the normalised differential cross-section measured in the fiducial phase-space as a function of the $|\Delta\eta|$ between the two leptons in the electron-muon channel. The individual systematic uncertainties are symmetrized before deriving the correlation matrix.

The statistical correlation matrix of all the absolute differential cross-sections measured in the fiducial phase-space in the electron-muon channel.

The statistical correlation matrix of all the normalised differential cross-sections measured in the fiducial phase-space in the electron-muon channel.

Fiducial region definition.

Observed and expected 95% confidence-level limits on sigma x BR as a function of mZ`.

Cutflow Table for 1 TeV, 2 TeV, 4 TeV MC signal samples and Data. Cut0: $\mathcal{L} \times \sigma \times B(Z`\to H\gamma)\times B(H \to b\bar{b})$; Cut1: Preselection; Cut2: Photon requirement; Cut3: Large-$R$ jet requirement; Cut4: Large-$R$ jet mass optimization; Cut5: Double $b$-tagged; Cut5.1: $p_T$ optimization(Double $b$-tagged); Cut6: Single $b$-tagged; Cut6.1: $p_T$ optimization(Single $b$-tagged).

Distribution of the resonance mass in the c-tag Signal Region of the $ ce$ channel for the post-fit background, the observed data, and the expected signal with $m_{LQ} = 1$ TeV.

Distribution of the resonance mass in the b-tag Signal Region of the $ ce$ channel for the post-fit background, the observed data, and the expected signal with $m_{LQ} = 1$ TeV.

Distribution of the resonance mass in the untagged Signal Region of the $ c\mu$ channel for the post-fit background, the observed data, and the expected signal with $m_{LQ} = 1$ TeV.

Distribution of the resonance mass in the c-tag Signal Region of the $ c\mu$ channel for the post-fit background, the observed data, and the expected signal with $m_{LQ} = 1$ TeV.

Distribution of the resonance mass in the b-tag Signal Region of the $ c\mu$ channel for the post-fit background, the observed data, and the expected signal with $m_{LQ} = 1$ TeV.

Distribution of the resonance mass in the 0-tag Signal Region of the $ be$ channel for the post-fit background, the observed data, and the expected signal with $m_{LQ} = 1$ TeV.

Distribution of the resonance mass in the 1-tag Signal Region of the $ be$ channel for the post-fit background, the observed data, and the expected signal with $m_{LQ} = 1$ TeV.

Distribution of the resonance mass in the 2-tag Signal Region of the $ be$ channel for the post-fit background, the observed data, and the expected signal with $m_{LQ} = 1$ TeV.

Distribution of the resonance mass in the 0-tag Signal Region of the $ b\mu$ channel for the post-fit background, the observed data, and the expected signal with $m_{LQ} = 1$ TeV.

Distribution of the resonance mass in the 1-tag Signal Region of the $ b\mu$ channel for the post-fit background, the observed data, and the expected signal with $m_{LQ} = 1$ TeV.

Distribution of the resonance mass in the 2-tag Signal Region of the $ b\mu$ channel for the post-fit background, the observed data, and the expected signal with $m_{LQ} = 1$ TeV.

The observed and expected limits on the leptoquark pair production cross-section at 95% CL for $\mathcal{B}=1$ into electrons, shown as a function of $m_{LQ}$ for the $qe$ channel.

The observed and expected limits on the leptoquark pair production cross-section at 95% CL for $\mathcal{B}=1$ into muons, shown as a function of $m_{LQ}$ for the $q\mu$ channel.

The observed and expected limits on the leptoquark pair production cross-section at 95% CL for $\mathcal{B}=1$ into electrons, shown as a function of $m_{LQ}$ for the $ce$ channel.

The observed and expected limits on the leptoquark pair production cross-section at 95% CL for $\mathcal{B}=1$ into muons, shown as a function of $m_{LQ}$ for the $c\mu$ channel.

The observed and expected limits on the leptoquark pair production cross-section at 95% CL for $\mathcal{B}=1$ into electrons, shown as a function of $m_{LQ}$ for the $be$ channel.

The observed and expected limits on the leptoquark pair production cross-section at 95% CL for $\mathcal{B}=1$ into muons, shown as a function of $m_{LQ}$ for the $b\mu$ channel.

The observed and expected limits on the leptoquark branching ratio at 95% CL, shown as a function of $m_{LQ}$ for the $qe$ channel.

The observed and expected limits on the leptoquark branching ratio at 95% CL, shown as a function of $m_{LQ}$ for the $q\mu$ channel.

The observed and expected limits on the leptoquark branching ratio at 95% CL, shown as a function of $m_{LQ}$ for the $ce$ channel.

The observed and expected limits on the leptoquark branching ratio at 95% CL, shown as a function of $m_{LQ}$ for the $c\mu$ channel.

The observed and expected limits on the leptoquark branching ratio at 95% CL, shown as a function of $m_{LQ}$ for the $be$ channel.

The observed and expected limits on the leptoquark branching ratio at 95% CL, shown as a function of $m_{LQ}$ for the $b\mu$ channel.

The signal selection efficiency x acceptance summed over all signal regions, for all masses and LQ decay channels considered.

The observed and expected limits for all masses and LQ decay channels considered.

Cutflow Table in the electron channel, considering signal samples with LQ mass of 1 TeV.

Cutflow Table in the muon channel, considering signal samples with LQ mass of 1 TeV.

Observed and expected 95% credibility-level upper limits on the cross-section times branching ratio for the simplified dark-matter model. The table also shows the corresponding $1\sigma$ and $2\sigma$ bands for the expected limits. The limits are calculated using jets in events with at least one isolated lepton ($e$ or $\mu$) with $p_\text{T}^\ell \ge 60$ GeV.

The 95% CL upper limits on the cross-section times acceptance times efficiency times branching ratio, as a function of the resonance mass, $m_X$. The limits are calculated using jets in events with at least one isolated lepton ($e$ or $\mu$) with $p_\text{T}^\ell \ge 60$ GeV. The limits are shown for a generic Gaussian signal with the width $\sigma_{X}/m_{X} =0$. The table also shows the expected limits and the corresponding $1\sigma$ and $2\sigma$ bands. The tabulated data has been used in Figure 04 of the publication.

The 95% CL upper limits on the cross-section times acceptance times efficiency times branching ratio, as a function of the resonance mass, $m_X$. The limits are calculated using jets in events with at least one isolated lepton ($e$ or $\mu$) with $p_\text{T}^\ell \ge 60$ GeV. The limits are shown for a generic Gaussian signal with the width $\sigma_{X}/m_{X} =0.05$. The table also shows the expected limits and the corresponding $1\sigma$ and $2\sigma$ bands. The tabulated data has been used in Figure 04 of the publication.

The 95% CL upper limits on the cross-section times acceptance times efficiency times branching ratio, as a function of the resonance mass, $m_X$. The limits are calculated using jets in events with at least one isolated lepton ($e$ or $\mu$) with $p_\text{T}^\ell \ge 60$ GeV. The limits are shown for a generic Gaussian signal with the width $\sigma_{X}/m_{X} =0.10$. The table also shows the expected limits and the corresponding $1\sigma$ and $2\sigma$ bands. The tabulated data has been used in Figure 04 of the publication.

The 95% CL upper limits on the cross-section times acceptance times efficiency times branching ratio, as a function of the resonance mass, $m_X$. The limits are calculated using jets in events with at least one isolated lepton ($e$ or $\mu$) with $p_\text{T}^\ell \ge 60$ GeV. The limits are shown for a generic Gaussian signal with the width $\sigma_{X}/m_{X} =0.15$. The table also shows the expected limits and the corresponding $1\sigma$ and $2\sigma$ bands. The tabulated data has been used in Figure 04 of the publication.

Truth-level cross-sections ($\sigma$), acceptances ($A$), detector efficiencies ($\epsilon$) and the detector-level visible cross-sections ($\sigma^{vis}$) for different masses of the $\rho_T$ model. Only statistical uncertainties on the acceptances and efficiencies are shown. The acceptance is defined by requiring jets with $m_{jj} >216$ GeV and $|\eta|<2.4$, as well as isolated leptons ($e$ or $\mu$) with $p_\text{T}^\ell \ge 60$ GeV and $|\eta^\mu|<2.5$ ($|\eta^e|<2.47$, excluding of the gap region 1.37--1.52) for muons (electrons). The tabulated data has been used in Figure 05a of the publication.

Truth-level cross-sections ($\sigma$), acceptances ($A$), detector efficiencies ($\epsilon$) and the detector-level visible cross-sections ($\sigma^{vis}$) for different masses of the $W'/Z'$ SSM model. Only statistical uncertainties on the acceptances and efficiencies are shown. The acceptance is defined by requiring jets with $m_{jj} >216$ GeV and $|\eta|<2.4$, as well as leptons with $p_\text{T}^\ell \ge 60$ GeV and $|\eta^\mu|<2.5$ ($|\eta^e|<2.47$ with exclusion of the gap region 1.37-1.52) for muons (electrons). The tabulated data has been used in Figure 05b of the publication.

Truth-level cross-sections ($\sigma$), acceptances ($A$), detector efficiencies ($\epsilon$) and the detector-level visible cross-sections ($\sigma^{vis}$) for different masses of the $H^+$ model for $\tan \beta=1$. Only statistical uncertainties on the acceptances and efficiencies are shown. The acceptance correction was calculated assuming $\tan \beta=1$ and the narrow-width approximation. The acceptance is defined by requiring jets with $m_{jj} >216$ GeV and $|\eta|<2.4$, as well as leptons with $p_\text{T}^\ell \ge 60$ GeV and $|\eta^\mu|<2.5$ ($|\eta^e|<2.47$ with exclusion of the gap region 1.37-1.52) for muons (electrons). The tabulated data has been used in Figure 05c of the publication.

Truth-level cross-sections ($\sigma$), acceptances ($A$), detector efficiencies ($\epsilon$) and the detector-level visible cross-sections ($\sigma^{vis}$) for different masses of the $Z'$ (DM) model. Only statistical uncertainties on the acceptances and efficiencies are shown. The acceptance is defined by requiring jets with $m_{jj} > 216$ GeV and $|\eta|<2.4$, as well as leptons with $p_\text{T}^\ell \ge 60$ GeV and $|\eta^\mu|<2.5$ ($|\eta^e|<2.47$ with exclusion of the gap region 1.37-1.52) for muons (electrons). The tabulated data has been used in Figure 05d of the publication.

A comparison of the measured cell $E_T^{miss}$ distribution with that predicted by the two-component model for two pile-up scenarios. The magenta points extend the measured distribution using L1 $E_T^{miss} > 30\,$GeV and L1 $E_T^{miss} > 50\,$GeV data. The red curve is the prediction from the calorimeter-resolution part of the model. The green curve is the high $E_T^{miss}$ tail's probability distribution for the mean number of $pp$ interactions $\mu$ in each figure. The blue curve is the full model prediction computed by combining the $E_T^{miss}$ from these two individual sources shown in red and green, each calculated for $\mu=\langle\mu\rangle$. The black points show the unbiased $E_T^{miss}$ distribution measured in data. Corresponds to a prediction for $\langle\mu\rangle = 25$.

A comparison of the measured cell $E_T^{miss}$ distribution with that predicted by the two-component model for two pile-up scenarios. The magenta points extend the measured distribution using L1 $E_T^{miss} > 30\,$GeV and L1 $E_T^{miss} > 50\,$GeV data. The red curve is the prediction from the calorimeter-resolution part of the model. The green curve is the high $E_T^{miss}$ tail's probability distribution for the mean number of $pp$ interactions $\mu$ in each figure. The blue curve is the full model prediction computed by combining the $E_T^{miss}$ from these two individual sources shown in red and green, each calculated for $\mu=\langle\mu\rangle$. The black points show the unbiased $E_T^{miss}$ distribution measured in data. Corresponds to a prediction for $\langle\mu\rangle = 25$.

A comparison of the measured cell $E_T^{miss}$ distribution with that predicted by the two-component model for two pile-up scenarios. The magenta points extend the measured distribution using L1 $E_T^{miss} > 30\,$GeV and L1 $E_T^{miss} > 50\,$GeV data. The red curve is the prediction from the calorimeter-resolution part of the model. The green curve is the high $E_T^{miss}$ tail's probability distribution for the mean number of $pp$ interactions $\mu$ in each figure. The blue curve is the full model prediction computed by combining the $E_T^{miss}$ from these two individual sources shown in red and green, each calculated for $\mu=\langle\mu\rangle$. The black points show the unbiased $E_T^{miss}$ distribution measured in data. Corresponds to a prediction for $\langle\mu\rangle = 55$.

A comparison of the measured cell $E_T^{miss}$ distribution with that predicted by the two-component model for two pile-up scenarios. The magenta points extend the measured distribution using L1 $E_T^{miss} > 30\,$GeV and L1 $E_T^{miss} > 50\,$GeV data. The red curve is the prediction from the calorimeter-resolution part of the model. The green curve is the high $E_T^{miss}$ tail's probability distribution for the mean number of $pp$ interactions $\mu$ in each figure. The blue curve is the full model prediction computed by combining the $E_T^{miss}$ from these two individual sources shown in red and green, each calculated for $\mu=\langle\mu\rangle$. The black points show the unbiased $E_T^{miss}$ distribution measured in data. Corresponds to a prediction for $\langle\mu\rangle = 55$.

A comparison of the measured cell $E_T^{miss}$ distribution with that predicted by the two-component model for two pile-up scenarios. The magenta points extend the measured distribution using L1 $E_T^{miss} > 30\,$GeV and L1 $E_T^{miss} > 50\,$GeV data. The red curve is the prediction from the calorimeter-resolution part of the model. The green curve is the high $E_T^{miss}$ tail's probability distribution for the mean number of $pp$ interactions $\mu$ in each figure. The blue curve is the full model prediction computed by combining the $E_T^{miss}$ from these two individual sources shown in red and green, each calculated for $\mu=\langle\mu\rangle$. The black points show the unbiased $E_T^{miss}$ distribution measured in data. Corresponds to a prediction for $\langle\mu\rangle = 55$.

A comparison of the measured cell $E_T^{miss}$ distribution with that predicted by the two-component model for two pile-up scenarios. The magenta points extend the measured distribution using L1 $E_T^{miss} > 30\,$GeV and L1 $E_T^{miss} > 50\,$GeV data. The red curve is the prediction from the calorimeter-resolution part of the model. The green curve is the high $E_T^{miss}$ tail's probability distribution for the mean number of $pp$ interactions $\mu$ in each figure. The blue curve is the full model prediction computed by combining the $E_T^{miss}$ from these two individual sources shown in red and green, each calculated for $\mu=\langle\mu\rangle$. The black points show the unbiased $E_T^{miss}$ distribution measured in data. Corresponds to a prediction for $\langle\mu\rangle = 55$.

A comparison of the measured cell $E_T^{miss}$ distribution with that predicted by the two-component model for two pile-up scenarios. The magenta points extend the measured distribution using L1 $E_T^{miss} > 30\,$GeV and L1 $E_T^{miss} > 50\,$GeV data. The red curve is the prediction from the calorimeter-resolution part of the model. The green curve is the high $E_T^{miss}$ tail's probability distribution for the mean number of $pp$ interactions $\mu$ in each figure. The blue curve is the full model prediction computed by combining the $E_T^{miss}$ from these two individual sources shown in red and green, each calculated for $\mu=\langle\mu\rangle$. The black points show the unbiased $E_T^{miss}$ distribution measured in data. Corresponds to a prediction for $\langle\mu\rangle = 55$.

The $E_T^{miss}$ model predicted trigger rate as a function of $\mu$ for the cell $E_T^{miss}$ algorithm with a threshold of $80\,$GeV and $120\,$GeV, assuming no additional pile-up mitigation.

The L1 $E_T^{miss}$ trigger efficiency, shown as a function of $p_T(\mu\mu)$ in $Z\rightarrow\mu\mu$ events.

The efficiencies in the plot are shown for events satisfying a $Z\rightarrow\mu\mu$ selection and with $p_T(\mu\mu)$ larger than $150\,$GeV vs pile-up for each of the four years of data taking.

The L1 $E_T^{miss}$ trigger rate as a function of $\langle\mu\rangle$ for runs in three different periods ($\alpha$, $\beta$, $\gamma$) in the year 2017.

The L1 $E_T^{miss}$ trigger rate as a function of $\langle\mu\rangle$ for runs in three different periods ($\alpha$, $\beta$, $\gamma$) in the year 2017.

The L1 $E_T^{miss}$ trigger rate as a function of $\langle\mu\rangle$ for runs in three different periods ($\alpha$, $\beta$, $\gamma$) in the year 2017.

The L1 $E_T^{miss}$ trigger efficiency is shown as a function of mean pile-up for events satisfying a $Z\rightarrow\mu\mu$ selection and with $p_T(\mu\mu)$ larger than $150\,$GeV in three periods during the year 2017.

Background acceptance vs signal efficiency for each of four individual HLT $E_T^{miss}$ algorithms for a $Z\rightarrow\mu\mu$ selection with $p_T(\mu\mu) > 175\,$GeV for data recorded in the year 2017. The diamond indicates the performance of the pufit $E_T^{miss}>110\,$GeV trigger. $0\leq\langle\mu\rangle < 20$.

Background acceptance vs signal efficiency for each of four individual HLT $E_T^{miss}$ algorithms for a $Z\rightarrow\mu\mu$ selection with $p_T(\mu\mu) > 175\,$GeV for data recorded in the year 2017. The diamond indicates the performance of the pufit $E_T^{miss}>110\,$GeV trigger. $0\leq\langle\mu\rangle < 20$.

Background acceptance vs signal efficiency for each of four individual HLT $E_T^{miss}$ algorithms for a $Z\rightarrow\mu\mu$ selection with $p_T(\mu\mu) > 175\,$GeV for data recorded in the year 2017. The diamond indicates the performance of the pufit $E_T^{miss}>110\,$GeV trigger. $0\leq\langle\mu\rangle < 20$.

Background acceptance vs signal efficiency for each of four individual HLT $E_T^{miss}$ algorithms for a $Z\rightarrow\mu\mu$ selection with $p_T(\mu\mu) > 175\,$GeV for data recorded in the year 2017. The diamond indicates the performance of the pufit $E_T^{miss}>110\,$GeV trigger. $0\leq\langle\mu\rangle < 20$.

Background acceptance vs signal efficiency for each of four individual HLT $E_T^{miss}$ algorithms for a $Z\rightarrow\mu\mu$ selection with $p_T(\mu\mu) > 175\,$GeV for data recorded in the year 2017. The diamond indicates the performance of the pufit $E_T^{miss}>110\,$GeV trigger. $0\leq\langle\mu\rangle < 20$.

Background acceptance vs signal efficiency for each of four individual HLT $E_T^{miss}$ algorithms for a $Z\rightarrow\mu\mu$ selection with $p_T(\mu\mu) > 175\,$GeV for data recorded in the year 2017. The diamond indicates the performance of the pufit $E_T^{miss}>110\,$GeV trigger. $20\leq\langle\mu\rangle < 30$.

Background acceptance vs signal efficiency for each of four individual HLT $E_T^{miss}$ algorithms for a $Z\rightarrow\mu\mu$ selection with $p_T(\mu\mu) > 175\,$GeV for data recorded in the year 2017. The diamond indicates the performance of the pufit $E_T^{miss}>110\,$GeV trigger. $20\leq\langle\mu\rangle < 30$.

Background acceptance vs signal efficiency for each of four individual HLT $E_T^{miss}$ algorithms for a $Z\rightarrow\mu\mu$ selection with $p_T(\mu\mu) > 175\,$GeV for data recorded in the year 2017. The diamond indicates the performance of the pufit $E_T^{miss}>110\,$GeV trigger. $20\leq\langle\mu\rangle < 30$.

Background acceptance vs signal efficiency for each of four individual HLT $E_T^{miss}$ algorithms for a $Z\rightarrow\mu\mu$ selection with $p_T(\mu\mu) > 175\,$GeV for data recorded in the year 2017. The diamond indicates the performance of the pufit $E_T^{miss}>110\,$GeV trigger. $20\leq\langle\mu\rangle < 30$.

Background acceptance vs signal efficiency for each of four individual HLT $E_T^{miss}$ algorithms for a $Z\rightarrow\mu\mu$ selection with $p_T(\mu\mu) > 175\,$GeV for data recorded in the year 2017. The diamond indicates the performance of the pufit $E_T^{miss}>110\,$GeV trigger. $20\leq\langle\mu\rangle < 30$.

Background acceptance vs signal efficiency for each of four individual HLT $E_T^{miss}$ algorithms for a $Z\rightarrow\mu\mu$ selection with $p_T(\mu\mu) > 175\,$GeV for data recorded in the year 2017. The diamond indicates the performance of the pufit $E_T^{miss}>110\,$GeV trigger. $30\leq\langle\mu\rangle < 40$.

Background acceptance vs signal efficiency for each of four individual HLT $E_T^{miss}$ algorithms for a $Z\rightarrow\mu\mu$ selection with $p_T(\mu\mu) > 175\,$GeV for data recorded in the year 2017. The diamond indicates the performance of the pufit $E_T^{miss}>110\,$GeV trigger. $30\leq\langle\mu\rangle < 40$.

Background acceptance vs signal efficiency for each of four individual HLT $E_T^{miss}$ algorithms for a $Z\rightarrow\mu\mu$ selection with $p_T(\mu\mu) > 175\,$GeV for data recorded in the year 2017. The diamond indicates the performance of the pufit $E_T^{miss}>110\,$GeV trigger. $30\leq\langle\mu\rangle < 40$.

Background acceptance vs signal efficiency for each of four individual HLT $E_T^{miss}$ algorithms for a $Z\rightarrow\mu\mu$ selection with $p_T(\mu\mu) > 175\,$GeV for data recorded in the year 2017. The diamond indicates the performance of the pufit $E_T^{miss}>110\,$GeV trigger. $30\leq\langle\mu\rangle < 40$.

Background acceptance vs signal efficiency for each of four individual HLT $E_T^{miss}$ algorithms for a $Z\rightarrow\mu\mu$ selection with $p_T(\mu\mu) > 175\,$GeV for data recorded in the year 2017. The diamond indicates the performance of the pufit $E_T^{miss}>110\,$GeV trigger. $30\leq\langle\mu\rangle < 40$.

Background acceptance vs signal efficiency for each of four individual HLT $E_T^{miss}$ algorithms for a $Z\rightarrow\mu\mu$ selection with $p_T(\mu\mu) > 175\,$GeV for data recorded in the year 2017. The diamond indicates the performance of the pufit $E_T^{miss}>110\,$GeV trigger. $40\leq\langle\mu\rangle$.

Background acceptance vs signal efficiency for each of four individual HLT $E_T^{miss}$ algorithms for a $Z\rightarrow\mu\mu$ selection with $p_T(\mu\mu) > 175\,$GeV for data recorded in the year 2017. The diamond indicates the performance of the pufit $E_T^{miss}>110\,$GeV trigger. $40\leq\langle\mu\rangle$.

Background acceptance vs signal efficiency for each of four individual HLT $E_T^{miss}$ algorithms for a $Z\rightarrow\mu\mu$ selection with $p_T(\mu\mu) > 175\,$GeV for data recorded in the year 2017. The diamond indicates the performance of the pufit $E_T^{miss}>110\,$GeV trigger. $40\leq\langle\mu\rangle$.

Background acceptance vs signal efficiency for each of four individual HLT $E_T^{miss}$ algorithms for a $Z\rightarrow\mu\mu$ selection with $p_T(\mu\mu) > 175\,$GeV for data recorded in the year 2017. The diamond indicates the performance of the pufit $E_T^{miss}>110\,$GeV trigger. $40\leq\langle\mu\rangle$.

Background acceptance vs signal efficiency for each of four individual HLT $E_T^{miss}$ algorithms for a $Z\rightarrow\mu\mu$ selection with $p_T(\mu\mu) > 175\,$GeV for data recorded in the year 2017. The diamond indicates the performance of the pufit $E_T^{miss}>110\,$GeV trigger. $40\leq\langle\mu\rangle$.

Relative background acceptance fraction vs. relative efficiency for two different $p_T(\mu\mu)>150\,$GeV threshold for data recorded in the year 2018. Two of the curves show the performance of the stand-alone cell algorithm and the stand-alone pufit algorithm. The other two show combined algorithms each formed by requiring that the event satisfy both a fixed threshold (either 65 GeV or 70 GeV as shown in the legend) for the cell algorithm and a pufit threshold which varies along the curve. In each plot the background acceptance fractions and the efficiencies are relative to those of the pufit $E_T^{miss} > 110\,$GeV trigger and thus can be greater than one. The diamond indicates the performance of the pufit $E_T^{miss}>110\,$GeV trigger while the cross indicates the performance of the combined (pufit $E_T^{miss}>110\,$GeV and cell $E_T^{miss}>50\,$GeV) trigger.

Relative background acceptance fraction vs. relative efficiency for two different $p_T(\mu\mu)>150\,$GeV threshold for data recorded in the year 2018. Two of the curves show the performance of the stand-alone cell algorithm and the stand-alone pufit algorithm. The other two show combined algorithms each formed by requiring that the event satisfy both a fixed threshold (either 65 GeV or 70 GeV as shown in the legend) for the cell algorithm and a pufit threshold which varies along the curve. In each plot the background acceptance fractions and the efficiencies are relative to those of the pufit $E_T^{miss} > 110\,$GeV trigger and thus can be greater than one. The diamond indicates the performance of the pufit $E_T^{miss}>110\,$GeV trigger while the cross indicates the performance of the combined (pufit $E_T^{miss}>110\,$GeV and cell $E_T^{miss}>50\,$GeV) trigger.

Relative background acceptance fraction vs. relative efficiency for two different $p_T(\mu\mu)>150\,$GeV threshold for data recorded in the year 2018. Two of the curves show the performance of the stand-alone cell algorithm and the stand-alone pufit algorithm. The other two show combined algorithms each formed by requiring that the event satisfy both a fixed threshold (either 65 GeV or 70 GeV as shown in the legend) for the cell algorithm and a pufit threshold which varies along the curve. In each plot the background acceptance fractions and the efficiencies are relative to those of the pufit $E_T^{miss} > 110\,$GeV trigger and thus can be greater than one. The diamond indicates the performance of the pufit $E_T^{miss}>110\,$GeV trigger while the cross indicates the performance of the combined (pufit $E_T^{miss}>110\,$GeV and cell $E_T^{miss}>50\,$GeV) trigger.

Relative background acceptance fraction vs. relative efficiency for two different $p_T(\mu\mu)>150\,$GeV threshold for data recorded in the year 2018. Two of the curves show the performance of the stand-alone cell algorithm and the stand-alone pufit algorithm. The other two show combined algorithms each formed by requiring that the event satisfy both a fixed threshold (either 65 GeV or 70 GeV as shown in the legend) for the cell algorithm and a pufit threshold which varies along the curve. In each plot the background acceptance fractions and the efficiencies are relative to those of the pufit $E_T^{miss} > 110\,$GeV trigger and thus can be greater than one. The diamond indicates the performance of the pufit $E_T^{miss}>110\,$GeV trigger while the cross indicates the performance of the combined (pufit $E_T^{miss}>110\,$GeV and cell $E_T^{miss}>50\,$GeV) trigger.

Relative background acceptance fraction vs. relative efficiency for two different $p_T(\mu\mu)>150\,$GeV threshold for data recorded in the year 2018. Two of the curves show the performance of the stand-alone cell algorithm and the stand-alone pufit algorithm. The other two show combined algorithms each formed by requiring that the event satisfy both a fixed threshold (either 65 GeV or 70 GeV as shown in the legend) for the cell algorithm and a pufit threshold which varies along the curve. In each plot the background acceptance fractions and the efficiencies are relative to those of the pufit $E_T^{miss} > 110\,$GeV trigger and thus can be greater than one. The diamond indicates the performance of the pufit $E_T^{miss}>110\,$GeV trigger while the cross indicates the performance of the combined (pufit $E_T^{miss}>110\,$GeV and cell $E_T^{miss}>50\,$GeV) trigger.

Relative background acceptance fraction vs. relative efficiency for two different $p_T(\mu\mu)>150\,$GeV threshold for data recorded in the year 2018. Two of the curves show the performance of the stand-alone cell algorithm and the stand-alone pufit algorithm. The other two show combined algorithms each formed by requiring that the event satisfy both a fixed threshold (either 65 GeV or 70 GeV as shown in the legend) for the cell algorithm and a pufit threshold which varies along the curve. In each plot the background acceptance fractions and the efficiencies are relative to those of the pufit $E_T^{miss} > 110\,$GeV trigger and thus can be greater than one. The diamond indicates the performance of the pufit $E_T^{miss}>110\,$GeV trigger while the cross indicates the performance of the combined (pufit $E_T^{miss}>110\,$GeV and cell $E_T^{miss}>50\,$GeV) trigger.

Relative background acceptance fraction vs. relative efficiency for two different $p_T(\mu\mu)>150\,$GeV threshold for data recorded in the year 2018. Two of the curves show the performance of the stand-alone cell algorithm and the stand-alone pufit algorithm. The other two show combined algorithms each formed by requiring that the event satisfy both a fixed threshold (either 65 GeV or 70 GeV as shown in the legend) for the cell algorithm and a pufit threshold which varies along the curve. In each plot the background acceptance fractions and the efficiencies are relative to those of the pufit $E_T^{miss} > 110\,$GeV trigger and thus can be greater than one. The diamond indicates the performance of the pufit $E_T^{miss}>110\,$GeV trigger while the cross indicates the performance of the combined (pufit $E_T^{miss}>110\,$GeV and cell $E_T^{miss}>50\,$GeV) trigger.

Relative background acceptance fraction vs. relative efficiency for two different $p_T(\mu\mu)>150\,$GeV threshold for data recorded in the year 2018. Two of the curves show the performance of the stand-alone cell algorithm and the stand-alone pufit algorithm. The other two show combined algorithms each formed by requiring that the event satisfy both a fixed threshold (either 65 GeV or 70 GeV as shown in the legend) for the cell algorithm and a pufit threshold which varies along the curve. In each plot the background acceptance fractions and the efficiencies are relative to those of the pufit $E_T^{miss} > 110\,$GeV trigger and thus can be greater than one. The diamond indicates the performance of the pufit $E_T^{miss}>110\,$GeV trigger while the cross indicates the performance of the combined (pufit $E_T^{miss}>110\,$GeV and cell $E_T^{miss}>50\,$GeV) trigger.

Relative background acceptance fraction vs. relative efficiency for two different $p_T(\mu\mu)>150\,$GeV threshold for data recorded in the year 2018. Two of the curves show the performance of the stand-alone cell algorithm and the stand-alone pufit algorithm. The other two show combined algorithms each formed by requiring that the event satisfy both a fixed threshold (either 65 GeV or 70 GeV as shown in the legend) for the cell algorithm and a pufit threshold which varies along the curve. In each plot the background acceptance fractions and the efficiencies are relative to those of the pufit $E_T^{miss} > 110\,$GeV trigger and thus can be greater than one. The diamond indicates the performance of the pufit $E_T^{miss}>110\,$GeV trigger while the cross indicates the performance of the combined (pufit $E_T^{miss}>110\,$GeV and cell $E_T^{miss}>50\,$GeV) trigger.

Relative background acceptance fraction vs. relative efficiency for two different $p_T(\mu\mu)>150\,$GeV threshold for data recorded in the year 2018. Two of the curves show the performance of the stand-alone cell algorithm and the stand-alone pufit algorithm. The other two show combined algorithms each formed by requiring that the event satisfy both a fixed threshold (either 65 GeV or 70 GeV as shown in the legend) for the cell algorithm and a pufit threshold which varies along the curve. In each plot the background acceptance fractions and the efficiencies are relative to those of the pufit $E_T^{miss} > 110\,$GeV trigger and thus can be greater than one. The diamond indicates the performance of the pufit $E_T^{miss}>110\,$GeV trigger while the cross indicates the performance of the combined (pufit $E_T^{miss}>110\,$GeV and cell $E_T^{miss}>50\,$GeV) trigger.

Relative background acceptance fraction vs. relative efficiency for two different $p_T(\mu\mu)>150\,$GeV threshold for data recorded in the year 2018. Two of the curves show the performance of the stand-alone cell algorithm and the stand-alone pufit algorithm. The other two show combined algorithms each formed by requiring that the event satisfy both a fixed threshold (either 65 GeV or 70 GeV as shown in the legend) for the cell algorithm and a pufit threshold which varies along the curve. In each plot the background acceptance fractions and the efficiencies are relative to those of the pufit $E_T^{miss} > 110\,$GeV trigger and thus can be greater than one. The diamond indicates the performance of the pufit $E_T^{miss}>110\,$GeV trigger while the cross indicates the performance of the combined (pufit $E_T^{miss}>110\,$GeV and cell $E_T^{miss}>50\,$GeV) trigger.

Relative background acceptance fraction vs. relative efficiency for two different $p_T(\mu\mu)>150\,$GeV threshold for data recorded in the year 2018. Two of the curves show the performance of the stand-alone cell algorithm and the stand-alone pufit algorithm. The other two show combined algorithms each formed by requiring that the event satisfy both a fixed threshold (either 65 GeV or 70 GeV as shown in the legend) for the cell algorithm and a pufit threshold which varies along the curve. In each plot the background acceptance fractions and the efficiencies are relative to those of the pufit $E_T^{miss} > 110\,$GeV trigger and thus can be greater than one. The diamond indicates the performance of the pufit $E_T^{miss}>110\,$GeV trigger while the cross indicates the performance of the combined (pufit $E_T^{miss}>110\,$GeV and cell $E_T^{miss}>50\,$GeV) trigger.

Turn-on efficiency curves are shown for $Z\rightarrow\mu\mu$ events for three algorithms: the cell algorithm alone, the pufit algorithm alone and the combined cell+pufit algorithm. The thresholds are set such that the algorithms have equal rates, and the data were recorded in the year 2018. Here the trigger efficiency is shown with respect to $p_T(\mu\mu)$.

Turn-on efficiency curves are shown for $Z\rightarrow\mu\mu$ events for three algorithms: the cell algorithm alone, the pufit algorithm alone and the combined cell+pufit algorithm. The thresholds are set such that the algorithms have equal rates, and the data were recorded in the year 2018. Here the trigger efficiency is shown with respect to the offline $E_T^{miss}$ calculation with muons treated as being invisible.

Efficiencies for $Z\rightarrow\mu\mu$ events are shown for the L1 $E_T^{miss}>50\,$GeV trigger and for the complete L1+HLT trigger chain that also requires pufit > 110 GeV. Each is shown as a function of $\langle\mu\rangle$ for a $p_T(\mu\mu)$ threshold of 150 GeV.

Efficiencies for $Z\rightarrow\mu\mu$ events are shown for the L1 $E_T^{miss}>50\,$GeV trigger and for the complete L1+HLT trigger chain that also requires pufit > 110 GeV. Each is shown as a function of $\langle\mu\rangle$ for an offline $E_T^{miss}$ threshold of 150 GeV.

Efficiencies for $Z\rightarrow\mu\mu$ events are shown for the L1 $E_T^{miss}>50\,$GeV trigger and for the complete L1+HLT trigger chain that also requires pufit > 110 GeV. Each is shown as a function of $\langle\mu\rangle$ for an $p_T(\mu\mu)$ threshold of 175 GeV.

Efficiencies for $Z\rightarrow\mu\mu$ events are shown for the L1 $E_T^{miss}>50\,$GeV trigger and for the complete L1+HLT trigger chain that also requires pufit > 110 GeV. Each is shown as a function of $\langle\mu\rangle$ for an offline $E_T^{miss}$ threshold of 175 GeV.

HLT_xe70_mht trigger output rate as a function of $\langle\mu\rangle$ shown for an example run in year 2015.

HLT_xe90_mht trigger output rate as a function of $\langle\mu\rangle$ shown for an example run in year 2016.

HLT_xe110_mht trigger output rate as a function of $\langle\mu\rangle$ shown for an example run in year 2016.

HLT_xe110_pufit trigger output rate as a function of $\langle\mu\rangle$ shown for an example run in year 2017. This trigger included an implicit requirement of cell $E_T^{miss}>50\,$GeV.

HLT_xe110_pufit_xe65 trigger output rate as a function of $\langle\mu\rangle$ shown for an example run in year 2018.

HLT_xe110_pufit_xe70 trigger output rate as a function of $\langle\mu\rangle$ shown for an example run in year 2018.

Full-chain trigger efficiencies for each year as a function of $p_T(\mu\mu)$. The efficiency corresponds to that of the lowest unprescaled trigger that is adjusted throughout each year.

Full-chain trigger efficiencies for 2015, 2016, 2017 and 2018 as a function of $\langle\mu\rangle$ for $p_T(\mu\mu)>150$ GeV. The efficiency corresponds to that of the lowest unprescaled trigger that is adjusted throughout the year.

Efficiencies for the first-level trigger L1XE50 and the combined L1+HLT trigger chain HLT_xe110_pufit_xe65_L1XE50 in data recorded in the year 2018 are shown as a function of $\langle\mu\rangle$ for $W\rightarrow e\nu$ and $Z\rightarrow\mu\mu$ selections with offline $E_T^{miss}>150\,$GeV.

Efficiencies for the first-level trigger L1XE50 and the combined L1+HLT trigger chain HLT_xe110_pufit_xe65_L1XE50 in data recorded in the year 2018 are shown as a function of $\langle\mu\rangle$ for $t\bar t$ with offline $E_T^{miss}>150\,$GeV.

Efficiencies for the first-level trigger L1XE50 and the combined L1+HLT trigger chain HLT_xe110_pufit_xe65_L1XE50 in data recorded in the year 2018 are shown as a function of $\langle\mu\rangle$ for $W\rightarrow e\nu$ and $Z\rightarrow\mu\mu$ selections with offline $E_T^{miss}>175\,$GeV.

Efficiencies for the first-level trigger L1XE50 and the combined L1+HLT trigger chain HLT_xe110_pufit_xe65_L1XE50 in data recorded in the year 2018 are shown as a function of $\langle\mu\rangle$ for $t\bar t$ with offline $E_T^{miss}>175\,$GeV.

The observed upper limits on the model cross-section in units of pb for simplified models with ${\tilde{\tau}}_L {\tilde{\tau}}_L$ only production. Three points at $M({\tilde{\chi}}^{0}_{1})=200GeV$ were removed from the plot but kept in the table because they overlapped with the plot's legend and are far from the exclusion contour.

The observed 95\% CL exclusion contours for the combined fit of SR-lowMass and SR-highMass for simplified models with combined ${\tilde{\tau}}^{+}_{R,L} {\tilde{\tau}}^{-}_{R,L}$ production.

The observed 95\% CL exclusion contours for the combined fit of SR-lowMass and SR-highMass for simplified models with combined ${\tilde{\tau}}^{+}_{R,L} {\tilde{\tau}}^{-}_{R,L}$ production.

The expected 95% CL exclusion contours for the combined fit of SR-lowMass and SR-highMass for simplified models with combined ${\tilde{\tau}}^{+}_{R,L} {\tilde{\tau}}^{-}_{R,L}$ production.

The expected 95% CL exclusion contours for the combined fit of SR-lowMass and SR-highMass for simplified models with combined ${\tilde{\tau}}^{+}_{R,L} {\tilde{\tau}}^{-}_{R,L}$ production.

The observed 95% CL exclusion contours for the combined fit of SR-lowMass and SR-highMass for simplified models with ${\tilde{\tau}}_L {\tilde{\tau}}_L$ only production.

The observed 95% CL exclusion contours for the combined fit of SR-lowMass and SR-highMass for simplified models with ${\tilde{\tau}}_L {\tilde{\tau}}_L$ only production.

The expected 95% CL exclusion contours for the combined fit of SR-lowMass and SR-highMass for simplified models with ${\tilde{\tau}}_L {\tilde{\tau}}_L$ only production.

The expected 95% CL exclusion contours for the combined fit of SR-lowMass and SR-highMass for simplified models with ${\tilde{\tau}}_L {\tilde{\tau}}_L$ only production.

Observed 95% CL exclusion limits for simplified models with direct stau pair production in SR-lowMass.

Observed 95% CL exclusion limits for simplified models with direct stau pair production in SR-lowMass.

Expected 95% CL exclusion limits for simplified models with direct stau pair production in SR-lowMass.

Expected 95% CL exclusion limits for simplified models with direct stau pair production in SR-lowMass.

Observed 95% CL exclusion limits for simplified models with direct stau pair production in SR-highMass.

Observed 95% CL exclusion limits for simplified models with direct stau pair production in SR-highMass.

Expected 95% CL exclusion limits for simplified models with direct stau pair production in SR-highMass.

Expected 95% CL exclusion limits for simplified models with direct stau pair production in SR-highMass.

Signal acceptance in SR highMass for combined stau final states

Signal acceptance in SR highMass for combined stau final states

Signal acceptance in SR lowMass for combined stau final states

Signal acceptance in SR lowMass for combined stau final states

Signal efficiency in SR highMass for combined stau final states

Signal efficiency in SR highMass for combined stau final states

Signal efficiency in SR lowMass for combined stau final states

Signal efficiency in SR lowMass for combined stau final states

Signal acceptance*efficiency in SR highMass for combined stau final states

Signal acceptance*efficiency in SR highMass for combined stau final states

Signal acceptance*efficiency in SR lowMass for combined stau final states

Signal acceptance*efficiency in SR lowMass for combined stau final states

Cutflow for two reference points (${\tilde{\tau}}^{+}_{R,L} {\tilde{\tau}}^{-}_{R,L}$ production) in SR. The column labelled $N_{weighted}$ shows the results including all correction factors applied to simulation, and is normalised to 139 fb$^{-1}$, while $N_{raw}$ in brackets shows the results for the generated number of events. The quoted uncertainties are statistical only. The "Generator filter" includes the requirements that two $\tau$ in the event have ${p}_{T} > 15$ GeV and $|\eta| <$ 2.6. The "Baseline Cut" includes the requirement of two baseline $\tau$ with a minimum value at 0.01 of the boosted decision tree discriminant (JetBDTSigTransMin $>$ 0.01) and ${p}_{T, \tau_{1}} > 50$ GeV and ${p}_{T, \tau_{2}} > 40$ GeV. At the step "Trigger & offline cuts", the following requirements are applied: the event is recorded using the asymmetric di-$\tau$ trigger (di-$\tau$ $E_{T}^{miss}$ trigger) in SR-lowMass (SR-highMass), and the lepton $p_{T}$ and $E_{T}^{miss}$ are required at plateau.

Cutflow for two reference points (${\tilde{\tau}}^{+}_{R,L} {\tilde{\tau}}^{-}_{R,L}$ production) in SR. The column labelled $N_{weighted}$ shows the results including all correction factors applied to simulation, and is normalised to 139 fb$^{-1}$, while $N_{raw}$ in brackets shows the results for the generated number of events. The quoted uncertainties are statistical only. The "Generator filter" includes the requirements that two $\tau$ in the event have ${p}_{T} > 15$ GeV and $|\eta| <$ 2.6. The "Baseline Cut" includes the requirement of two baseline $\tau$ with a minimum value at 0.01 of the boosted decision tree discriminant (JetBDTSigTransMin $>$ 0.01) and ${p}_{T, \tau_{1}} > 50$ GeV and ${p}_{T, \tau_{2}} > 40$ GeV. At the step "Trigger & offline cuts", the following requirements are applied: the event is recorded using the asymmetric di-$\tau$ trigger (di-$\tau$ $E_{T}^{miss}$ trigger) in SR-lowMass (SR-highMass), and the lepton $p_{T}$ and $E_{T}^{miss}$ are required at plateau.

Observed and expected numbers of events in the control and signal regions where all control and signal region bins are included as constraints in the likelihood. The expected event yields of SM processes are given after the background-only fit. The entries marked as "--" are negligible. The uncertainties correspond to the sum in quadrature of statistical and systematic uncertainties. The correlation of systematic uncertainties among control regions and among background processes is fully taken into account.

Observed and expected numbers of events in the control and signal regions where all control and signal region bins are included as constraints in the likelihood. The expected event yields of SM processes are given after the background-only fit. The entries marked as "--" are negligible. The uncertainties correspond to the sum in quadrature of statistical and systematic uncertainties. The correlation of systematic uncertainties among control regions and among background processes is fully taken into account.

The post-fit $m_{T2}$ distribution for SR-lowMass. The stacked histograms show the expected SM backgrounds. The multi-jet contribution is estimated from data using the ABCD method. The contributions of multi-jet and $W$+jets events are scaled with the corresponding normalization factors derived from the background-only fit. The hatched bands represent the sum in quadrature of systematic and statistical uncertainties of the total SM background. For illustration, the distributions from the SUSY reference points are also shown as dashed lines. The last bin includes the overflow events.

The post-fit $m_{T2}$ distribution for SR-lowMass. The stacked histograms show the expected SM backgrounds. The multi-jet contribution is estimated from data using the ABCD method. The contributions of multi-jet and $W$+jets events are scaled with the corresponding normalization factors derived from the background-only fit. The hatched bands represent the sum in quadrature of systematic and statistical uncertainties of the total SM background. For illustration, the distributions from the SUSY reference points are also shown as dashed lines. The last bin includes the overflow events.

The post-fit $m_{T2}$ distribution for SR-highMass. The stacked histograms show the expected SM backgrounds. The multi-jet contribution is estimated from data using the ABCD method. The contributions of multi-jet and $W$+jets events are scaled with the corresponding normalization factors derived from the background-only fit. The hatched bands represent the sum in quadrature of systematic and statistical uncertainties of the total SM background. For illustration, the distributions from the SUSY reference points are also shown as dashed lines. The last bin includes the overflow events.

The post-fit $m_{T2}$ distribution for SR-highMass. The stacked histograms show the expected SM backgrounds. The multi-jet contribution is estimated from data using the ABCD method. The contributions of multi-jet and $W$+jets events are scaled with the corresponding normalization factors derived from the background-only fit. The hatched bands represent the sum in quadrature of systematic and statistical uncertainties of the total SM background. For illustration, the distributions from the SUSY reference points are also shown as dashed lines. The last bin includes the overflow events.

The $m_{T2}$ post-fit distributions in the multi-jet background validation region VR-F (lowMass). The stacked histograms show the contribution of each relevant SM process. The multi-jet shape is taken from VR-E in the ABCD method and the normalization is determined by the transfer factor $T$ and rescaled by a correction factor determined by the fit. The hatched bands represent the combined statistical and systematic uncertainties in the sum of the SM backgrounds shown. For illustration, the distributions of the SUSY reference points are also shown as dashed lines.

The $m_{T2}$ post-fit distributions in the multi-jet background validation region VR-F (lowMass). The stacked histograms show the contribution of each relevant SM process. The multi-jet shape is taken from VR-E in the ABCD method and the normalization is determined by the transfer factor $T$ and rescaled by a correction factor determined by the fit. The hatched bands represent the combined statistical and systematic uncertainties in the sum of the SM backgrounds shown. For illustration, the distributions of the SUSY reference points are also shown as dashed lines.

The $m_{T2}$ post-fit distributions in the multi-jet background validation VR-F (highMass). The stacked histograms show the contribution of each relevant SM process. The multi-jet shape is taken from VR-E in the ABCD method and the normalization is determined by the transfer factor $T$ and rescaled by a correction factor determined by the fit. The hatched bands represent the combined statistical and systematic uncertainties in the sum of the SM backgrounds shown. For illustration, the distributions of the SUSY reference points are also shown as dashed lines.

The $m_{T2}$ post-fit distributions in the multi-jet background validation VR-F (highMass). The stacked histograms show the contribution of each relevant SM process. The multi-jet shape is taken from VR-E in the ABCD method and the normalization is determined by the transfer factor $T$ and rescaled by a correction factor determined by the fit. The hatched bands represent the combined statistical and systematic uncertainties in the sum of the SM backgrounds shown. For illustration, the distributions of the SUSY reference points are also shown as dashed lines.

The $E_{T}^{miss}$ post-fit distributions in the multi-jet background validation region VR-F (lowMass). The stacked histograms show the contribution of each relevant SM process. The multi-jet shape is taken from VR-E in the ABCD method and the normalization is determined by the transfer factor $T$ and rescaled by a correction factor determined by the fit. The hatched bands represent the combined statistical and systematic uncertainties in the sum of the SM backgrounds shown. For illustration, the distributions of the SUSY reference points are also shown as dashed lines.

The $E_{T}^{miss}$ post-fit distributions in the multi-jet background validation region VR-F (lowMass). The stacked histograms show the contribution of each relevant SM process. The multi-jet shape is taken from VR-E in the ABCD method and the normalization is determined by the transfer factor $T$ and rescaled by a correction factor determined by the fit. The hatched bands represent the combined statistical and systematic uncertainties in the sum of the SM backgrounds shown. For illustration, the distributions of the SUSY reference points are also shown as dashed lines.

The $E_{T}^{miss}$ post-fit distributions in the multi-jet background validation region VR-F (highMass). The stacked histograms show the contribution of each relevant SM process. The multi-jet shape is taken from VR-E in the ABCD method and the normalization is determined by the transfer factor $T$ and rescaled by a correction factor determined by the fit. The hatched bands represent the combined statistical and systematic uncertainties in the sum of the SM backgrounds shown. For illustration, the distributions of the SUSY reference points are also shown as dashed lines.

The $E_{T}^{miss}$ post-fit distributions in the multi-jet background validation region VR-F (highMass). The stacked histograms show the contribution of each relevant SM process. The multi-jet shape is taken from VR-E in the ABCD method and the normalization is determined by the transfer factor $T$ and rescaled by a correction factor determined by the fit. The hatched bands represent the combined statistical and systematic uncertainties in the sum of the SM backgrounds shown. For illustration, the distributions of the SUSY reference points are also shown as dashed lines.

The pre-fit $m_{T2}$ distribution in the $WCR$. The SM backgrounds other than multi-jet production are estimated from MC simulation. The multi-jet contribution is estimated from data using the OS--SS method. The hatched bands represent the combined statistical and systematic uncertainties of the total SM background. For illustration, the distributions of the SUSY reference points are also shown as dashed lines.

The pre-fit $m_{T2}$ distribution in the $WCR$. The SM backgrounds other than multi-jet production are estimated from MC simulation. The multi-jet contribution is estimated from data using the OS--SS method. The hatched bands represent the combined statistical and systematic uncertainties of the total SM background. For illustration, the distributions of the SUSY reference points are also shown as dashed lines.

The post-fit yields in the $WVR$, $TVRs$, $ZVRs$ and $VVVRs$. The SM backgrounds other than multi-jet production are estimated from MC simulation. The multi-jet contribution is negligible and is estimated from data using the ABCD method, using CRs obtained with the same technique used for the SRs. The hatched bands represent the combined statistical and systematic uncertainties of the total SM background. For illustration, the distributions of the SUSY reference points are also shown as dashed lines. The lower panels show the ratio of data to the SM background estimate.

The post-fit yields in the $WVR$, $TVRs$, $ZVRs$ and $VVVRs$. The SM backgrounds other than multi-jet production are estimated from MC simulation. The multi-jet contribution is negligible and is estimated from data using the ABCD method, using CRs obtained with the same technique used for the SRs. The hatched bands represent the combined statistical and systematic uncertainties of the total SM background. For illustration, the distributions of the SUSY reference points are also shown as dashed lines. The lower panels show the ratio of data to the SM background estimate.