MSSMvsSMRun2Legacy

Analysis specific software for MSSM (with SM categories) to be used within CombineHarvester

Setup software

wget https://raw.githubusercontent.com/KIT-CMS/MSSMvsSMRun2Legacy/master/scripts/checkout.sh
bash ./checkout.sh

After this, please put the datacard inputs into the shapes folder in the following structure:

<year>/<channel>/htt_<category>.inputs-mssm-vs-sm-Run<year>-<variable>.root

Cut-based SM analysis

Datacard creation

morph_parallel.py --output output --analysis sm --eras 2016,2017,2018 --category_list ${CMSSW_BASE}/src/CombineHarvester/MSSMvsSMRun2Legacy/input/control_region_categories.txt --variable mt_tot_puppi --parallel 5
morph_parallel.py --output output --analysis sm --eras 2016,2017,2018 --category_list ${CMSSW_BASE}/src/CombineHarvester/MSSMvsSMRun2Legacy/input/sm_categories.txt --variable m_sv_puppi --parallel 5

for era in 2016 2017 2018;
do
    mkdir -p output_sm/${era}/cmb/; rsync -av --progress output_sm/${era}/htt_*/*  output_sm/${era}/cmb/
done;
mkdir -p output_sm/combined/cmb/; rsync -av --progress output_sm/201?/htt_*/*  output_sm/combined/cmb/

Workspace creation

inclusive and stage 0:

ulimit -s unlimited
combineTool.py -M T2W -i output_sm/{2016,2017,2018,combined}/cmb --parallel 4 -o inclusive.root
combineTool.py -M T2W -i output_sm/{2016,2017,2018,combined}/cmb --parallel 4 -o stage0.root -P HiggsAnalysis.CombinedLimit.PhysicsModel:multiSignalModel --PO '"map=^.*/ggH125.?$:r_ggH[1,-9,11]"' --PO '"map=^.*/qqH125.?$:r_qqH[1,-9,11]"'

Signal strength scans

inclusive:

combineTool.py -M MultiDimFit -d output_sm/combined/cmb/inclusive.root --there --algo grid --robustFit 1 --X-rtd MINIMIZER_analytic --cminDefaultMinimizerStrategy 0 --cminDefaultMinimizerTolerance 0.01 --floatOtherPOIs 1 -t -1 --expectSignal 1 -v1 --alignEdges 1 --setParameterRanges r=0.85,1.15 --points 41 --split-points 1 -n .inclusive --job-mode condor --task-name r_1Dscan --dry-run

# after adapting the condor configuration, submit to batch system:
condor_submit condor_r_1Dscan.sub

stage 0:

# ggH
combineTool.py -M MultiDimFit -d output_sm/combined/cmb/stage0.root --there --algo grid --robustFit 1 --X-rtd MINIMIZER_analytic --cminDefaultMinimizerStrategy 0 --cminDefaultMinimizerTolerance 0.01 --floatOtherPOIs 1 -t -1 --expectSignal 1 -v1 --setParameters r_ggH=1,r_qqH=1 --redefineSignalPOIs r_ggH --alignEdges 1 --setParameterRanges r_ggH=0.65,1.35:r_qqH=0.3,1.7 --points 41 --split-points 1 -n .stage0_r_ggH --job-mode condor --task-name r_ggH_1Dscan --dry-run

# qqH
combineTool.py -M MultiDimFit -d output_sm/combined/cmb/stage0.root --there --algo grid --robustFit 1 --X-rtd MINIMIZER_analytic --cminDefaultMinimizerStrategy 0 --cminDefaultMinimizerTolerance 0.01 --floatOtherPOIs 1 -t -1 --expectSignal 1 -v1 --setParameters r_ggH=1,r_qqH=1 --redefineSignalPOIs r_qqH --alignEdges 1 --setParameterRanges r_qqH=0.65,1.35:r_ggH=0.3,1.7 --points 41 --split-points 1 -n .stage0_r_qqH --job-mode condor --task-name r_qqH_1Dscan --dry-run

# after adapting the condor configuration, submit to batch system:
condor_submit condor_r_ggH_1Dscan.sub
condor_submit condor_r_qqH_1Dscan.sub

# determining correlation ggH vs. qqH
combineTool.py -M FitDiagnostics -d output_sm/combined/cmb/stage0.root --there --robustHesse 1 --X-rtd MINIMIZER_analytic --cminDefaultMinimizerStrategy 0 --cminDefaultMinimizerTolerance 0.01 -t -1 --expectSignal 1 -v1 --setParameters r_ggH=1,r_qqH=1 --setParameterRanges r_qqH=0.65,1.35:r_ggH=0.65,1.35 -n .stage0 --task-name stage0_correlation

Collecting results & plotting likelihood scans:

hadd output_sm/combined/cmb/higgsCombine.inclusive.MultiDimFit.mH120.root output_sm/combined/cmb/higgsCombine.inclusive.POINTS*.root
hadd output_sm/combined/cmb/higgsCombine.stage0_r_ggH.MultiDimFit.mH120.root output_sm/combined/cmb/higgsCombine.stage0_r_ggH.POINTS*.root
hadd output_sm/combined/cmb/higgsCombine.stage0_r_qqH.MultiDimFit.mH120.root output_sm/combined/cmb/higgsCombine.stage0_r_qqH.POINTS*.root

plot1DScan.py output_sm/combined/cmb/higgsCombine.inclusive.MultiDimFit.mH120.root -o r_1Dscan
plot1DScan.py output_sm/combined/cmb/higgsCombine.stage0_r_ggH.MultiDimFit.mH120.root -o r_ggH_1Dscan --POI r_ggH
plot1DScan.py output_sm/combined/cmb/higgsCombine.stage0_r_qqH.MultiDimFit.mH120.root -o r_qqH_1Dscan --POI r_qqH

print_correlation.py output_sm/combined/cmb/fitDiagnostics.stage0.root r_ggH r_qqH

Prefit shapes

for card in output_MSSMvsSM_Run2_sm_nobtag_m_sv_puppi/*/cmb/combined.txt.cmb;
do
    PostFitShapesFromWorkspace -w ${card/combined.txt.cmb/stage0.root}  -o ${card/combined.txt.cmb/prefit_shapes.root} -d ${card}
done;

Model-independent MSSM analysis

Datacard creation

morph_parallel.py --output output --analysis mssm --eras 2016,2017,2018 --category_list ${CMSSW_BASE}/src/CombineHarvester/MSSMvsSMRun2Legacy/input/control_region_categories.txt --variable mt_tot_puppi --additional_arguments="--auto_rebin=1" --sm_gg_fractions ${CMSSW_BASE}/src/CombineHarvester/MSSMvsSMRun2Legacy/data/higgs_pt_v3.root --parallel 5
morph_parallel.py --output output --analysis mssm --eras 2016,2017,2018 --category_list ${CMSSW_BASE}/src/CombineHarvester/MSSMvsSMRun2Legacy/input/mssm_classic_categories.txt --variable mt_tot_puppi --additional_arguments="--auto_rebin=1" --sm_gg_fractions ${CMSSW_BASE}/src/CombineHarvester/MSSMvsSMRun2Legacy/data/higgs_pt_v3.root --parallel 5

for era in 2016 2017 2018;
do
    mkdir -p output_mssm/${era}/cmb/; rsync -av --progress output_mssm/${era}/htt_*/*  output_mssm/${era}/cmb/
done;
mkdir -p output_mssm/combined/cmb/; rsync -av --progress output_mssm/201?/htt_*/*  output_mssm/combined/cmb/

Workspace creation

ulimit -s unlimited
combineTool.py -M T2W -o "ws.root" -P HiggsAnalysis.CombinedLimit.PhysicsModel:multiSignalModel --PO '"map=^.*/ggh_(i|t|b).?$:r_ggH[0,0,200]"' --PO '"map=^.*/bbh$:r_bbH[0,0,200]"' -i output_mssm/{2016,2017,2018,combined}/cmb -m 110 --parallel 4

Model-independent CLs 95% limits (asymptotic, SM Higgs in background hypothesis)

bbH:

combineTool.py -m "110,120,130,140,160,180,200,250,300,350,400,450,500,600,700,800,900,1000,1200,1400,1600,1800,2000,2300,2600,2900,3200" -M AsymptoticLimits --rAbsAcc 0 --rRelAcc 0.0005 --boundlist ${CMSSW_BASE}/src/CombineHarvester/MSSMvsSMRun2Legacy/input/mssm_boundaries.json --setParameters r_ggH=0,r_bbH=0 --redefineSignalPOIs r_bbH -d output_mssm/combined/cmb/ws.root --there -n ".bbH" --job-mode condor --dry-run --task-name bbH_full_combined --X-rtd MINIMIZER_analytic --cminDefaultMinimizerStrategy 0 --cminDefaultMinimizerTolerance 0.01 -v 1

# After adaption of condor_bbH_full_combined.sub, submit to batch system:
condor_submit condor_bbH_full_combined.sub

ggH:

combineTool.py -m "110,120,130,140,160,180,200,250,300,350,400,450,500,600,700,800,900,1000,1200,1400,1600,1800,2000,2300,2600,2900,3200" -M AsymptoticLimits --rAbsAcc 0 --rRelAcc 0.0005 --boundlist ${CMSSW_BASE}/src/CombineHarvester/MSSMvsSMRun2Legacy/input/mssm_boundaries.json --setParameters r_ggH=0,r_bbH=0 --redefineSignalPOIs r_ggH -d output_mssm/combined/cmb/ws.root --there -n ".ggH" --job-mode condor --dry-run --task-name ggH_full_combined --X-rtd MINIMIZER_analytic --cminDefaultMinimizerStrategy 0 --cminDefaultMinimizerTolerance 0.01 -v 1

# After adaption of condor_ggH_full_combined.sub, submit to batch system:
condor_submit condor_ggH_full_combined.sub

Collecting limits:

for p in gg bb
do
    combineTool.py -M CollectLimits output_mssm/combined/cmb/higgsCombine.${p}H*.root --use-dirs -o output_mssm/combined/cmb/mssm_${p}H_combined.json
done

Plotting limits:

for p in gg bb
do
    plotMSSMLimits.py --cms-sub "Own Work" --title-right "137 fb^{-1} (13 TeV)" --process "${p}#phi" --y-axis-min 0.0001 --y-axis-max 1000.0 --show exp,obs output_mssm/combined/cmb/mssm_${p}H_combined_cmb.json  --output mssm_model-independent_combined_${p}H_cmb --logx --logy
done

Plotting of prefit shapes

Workspace creation:

In order to create prefit shapes from the available datacards, separate workspaces for the different analysis categories have to be created. This can be done with the following command

ulimit -s unlimited
combineTool.py -M T2W \
    -o "ws.root" \
    -P HiggsAnalysis.CombinedLimit.PhysicsModel:multiSignalModel \
    --PO '"map=^.*/ggh_(i|t|b).?$:r_ggH[0,0,200]"' \
    --PO '"map=^.*/bbh$:r_bbH[0,0,200]"' \
    -i output_mssm_classic/{2016,2017,2018}/htt_* \
    -m 110 \
    --parallel 5

Prefit shapes:

An exemplary command to extract prefit shapes from the created workspaces and datacards is

prefit_postfit_shapes_parallel.py \
    --datacard_pattern "output_mssm_classic/201?/htt_*/combined.txt.cmb" \
    --workspace_name ws.root \
    --output_name prefit_shapes.root \
    --freeze_arguments "--freeze MH=1200,r_ggH=0.1,r_bbH=0.1" \
    --parallel 5

hadd output_mssm_classic/combined/cmb/prefit_shapes.root output_mssm_classic/{2016,2017,2018}/htt_*/prefit_shapes.root

where the freeze arguments can be chosen freely.

Plotting of the created shapes:

Plotting scripts are provided in the plotting directory of this repository. The command to produce the plots of the prefit shapes created above is

plot_shapes_mssm_model_independent.sh $ERA output_mssm_classic/combined/cmb/prefit_shapes.root $OUTPUT_DIR et,mt,tt,em 1200

The last parameter given to the script is optional and will be the mass that is displayed in the legend of the plot. Currently the plotting is only possible for one experiment era a time. The output directory can be freely chosen and will be created if it is not yet present.

Running GoF tests

Workspace creation

The datacards from the usual datacard creation step can be directly used for the GoF tests. If the datacards are not available yet they can be produced via the run_model-indep_classic_limits.sh script. The script can then also be used to produce the necessary workspaces for the GoF tests. For the datacard and workspace creation the commands would be:

bash run_model-indep_classic_limits.sh auto initial
bash run_model-indep_classic_limits.sh auto ws-gof

where auto can be set to any tag value desired. The default when using auto as tag value is cmb_ind. These steps will create the workspaces separately for each category and era as well as their combinations in the directoy analysis/cmb_ind/datacards_bsm-model-indep if the tag value auto is used. The command used to produce the workspaces is

combineTool.py -M T2W -o "ws-gof.root" \
    -i analysis/cmb_ind/datacards_bsm-model-indep/*/{et,mt,tt,em,cmb}/ \
    --channel-masks \
    -m 125.0 --parallel 8

Running the GoF tests

A script is provided to run the GoF tests per category, per channel and year, per year and for the full combination. The script is located under gof/run_gof.sh. To run the GoF test on a single category the script should be invoked with to run the GoF test with 500 toys on a single core:

bash gof/run_gof.sh $ERA $CHANNEL $CATEGORY $PWD submit

To run the toy generation in parallel on 10 cores the argument submit can be changed to local. To run on the combined fit the easiest way is to call the script with the arguments:

bash gof/run_gof.sh combined cmb all $PWD local

Running channel compatibility checks

Workspace creation

The workspaces created to obtain the model-independent limits on the cross section times branching ratio of the production of additional neutral Higgs bosons can be used directly for the channel compatibility checks. To create the workspaces please refer to the corresponding section.

Running the channel compatibility checks

The command to obtain the Chi^2 like test statistics on data is:

combine -M ChannelCompatibilityCheck \
    ${WORKSPACE} \
    --X-rtd MINIMIZER_analytic --cminDefaultMinimizerStrategy 0 --cminDefaultMinimizerTolerance 0.01 --robustFit 1 --stepSize 0.005 \
    --setParameters r_ggH=0,r_bbH=0 --redefineSignalPOIs r_ggH --setParameterRange r_ggH=-200,80:r_bbH=-200,80 --rMin=-200 \
    -m ${MASS} \
    --saveFitResult \
    -n .Test.combined-cmb \
    -v 1

The command will calculate the test statistics by performing a fit with the common parameter of interest(POI) and a fit with separate POIs for each category in the provided workspace. To perform the channel compatibility check for a combination of categories the --group argument can be used to group the categories together, e.g. --group htt_em --group htt_et --group htt_mt --group htt_tt to perform the channel compatibility check per final state. Depending on the mass considered the parameter ranges for r_ggH and r_bbH and the step size might need to be adjusted to get the fit and the searches for the crossing points to converge. Note that the --saveFitResult is necessary to be able to plot the nominal and alternative fits afterwards.

The combine tool can be used to run the computation of the test statistics on toys. The command for this is then:

combineTool.py -M ChannelCompatibilityCheck \
    ${WORKSPACE} \
    --X-rtd MINIMIZER_analytic --cminDefaultMinimizerStrategy 0 --cminDefaultMinimizerTolerance 0.01 --robustFit 1 --stepSize 0.005 \ 
    --setParameters r_ggH=0,r_bbH=0 --redefineSignalPOIs r_ggH --setParameterRange r_ggH=-200,80:r_bbH=-200,80 --rMin=-200 \
    -m ${MASS} \
    -n .Test.combined-cmb \
    -v 0 \
    --dry-run -s 1230:1329:1 -t 5 --job-mode condor --task-name ChannelCompatibilityCheck.Test.combined-cmb.mH${MASS}

A faster runtime of the jobs but the same results with toys can probably achieved by adding the option --runMinos=false to the combine tool command as the uncertainties on the POIs are not of interest for the toys. Please note that this has not been tested so far. The splitting of the jobs is based on the range of the seeds given and can be adjusted to your needs.

As the outputs from the ChannelCompatibilityCheck method have the same outputs as the results from GoF tests, the collection and plotting scripts for the GoF tests can be used to extract the p-value of the tests and plot the distribution of the test statistics. The commands to do this are:

combineTool.py -M CollectGoodnessOfFit --input \
    higgsCombine.Test.combined-cmb.ChannelCompatibilityCheck.mH${MASS}.root higgsCombine.Test.combined-cmb.ChannelCompatibilityCheck.mH${MASS}.1230.root \
    higgsCombine.Test.combined-cmb.ChannelCompatibilityCheck.mH${MASS}.root higgsCombine.Test.combined-cmb.ChannelCompatibilityCheck.mH${MASS}.1231.root \
    ...
    higgsCombine.Test.combined-cmb.ChannelCompatibilityCheck.mH${MASS}.root higgsCombine.Test.combined-cmb.ChannelCompatibilityCheck.mH${MASS}.1329.root \
    --output ChannelCompatibilityCheck_combined_cmb-mH${MASS}.json

plotGof.py ChannelCompatibilityCheck_combined_cmb-mH${MASS}.json \
    --output ChannelCompatibilityCheck_combined_cmb-mH${MASS} \
    --statistic CCC \
    --mass ${MASS}.0 \
    --title-right="160 fb^{-1} (13 TeV)" --title-left="m_{#phi} = ${MASS} GeV"

To plot the nominal and the alternative fit results a script is provided that extracts the results for the POIs of both fits from the fit file that has been written by the fit on data. It is invoked via

python ${CMSSW_BASE}/src/CombineHarvester/MSSMvsSMRun2Legacy/plotting/plot_ccc.py \
    higgsCombine.Test.combined-cmb.ChannelCompatibilityCheck.mH${MASS}.root \
    -o ChannelCompatibilityCheck_FitResults_mH${MASS} \
    -p r_ggH \
    -r m0.012,0.01

where -r sets the x-range of the resulting plot.

Model-dependent MSSM analysis

In case of the calculation of model-dependent limits, several choices of signal models and categorization are available, denoted with an appropriate --analysis. The full BSM (in this case, MSSM) signal model consists of three neutral Higgs bosons, the light scalar h, the heavy scalar H, and the heavy pseudoscalar A. The production modes ggphi and bbphi are considered for all three Higgs bosons, whereas the qqphi production mode is accounted for only for the light scalar h.

The different analysis setups are listed in the following:

• mssm_vs_sm_classic: using classic categorization (similar to HIG-17-020) with the full signal model for the BSM prediction. The corresponding hypothesis is then compared with the SM prediction.
• mssm_vs_sm_classic_h125: same as above, but using for ggh the templates from the SM prediction, which are reweighted to the yield predicted by the MSSM scenario.
• mssm_vs_sm_heavy: using classic categorization (similar to HIG-17-020), but using only the heavy Higgs boson predictions H and A added on top of the background and SM prediction. This hypothesis is compared with the SM prediction.
• mssm_vs_sm: using SM categorization in addition to classic categories, with the full signal model for the BSM prediction. The corresponding hypothesis is then compared with the SM prediction. The BSM signal modelling is dependent on the categories:
  • for SM categories, only ggh and qqh are taken into account
  • high mass no-btag categories contain only bbH, bbA and ggH, ggA
  • btag categories & common control regions contain the full BSM signal model
• mssm_vs_sm_h125: same as above, but using for ggh the templates from the SM prediction, which are reweighted to the yield predicted by the MSSM scenario.
• mssm_vs_sm_CPV : here the considered Higgs bosons are H1 (SM-like), H2 and H3 instead of the usual h H and A. IN PROGRESS

Datacard creation for mssm_vs_sm_classic_h125

morph_parallel.py --output output --analysis mssm_vs_sm_classic_h125 --eras 2016,2017,2018 --category_list ${CMSSW_BASE}/src/CombineHarvester/MSSMvsSMRun2Legacy/input/control_region_categories.txt --variable mt_tot_puppi --parallel 5
morph_parallel.py --output output --analysis mssm_vs_sm_classic_h125 --eras 2016,2017,2018 --category_list ${CMSSW_BASE}/src/CombineHarvester/MSSMvsSMRun2Legacy/input/mssm_classic_categories.txt --variable mt_tot_puppi --parallel 5

mkdir -p output_mssm_vs_sm_classic_h125/combined/cmb/; rsync -av --progress output_mssm_vs_sm_classic_h125/201?/htt_*/*  output_mssm_vs_sm_classic_h125/combined/cmb/

Datacard creation for mssm_vs_sm_heavy

morph_parallel.py --output output --analysis mssm_vs_sm_heavy --eras 2016,2017,2018 --category_list ${CMSSW_BASE}/src/CombineHarvester/MSSMvsSMRun2Legacy/input/control_region_categories.txt --variable mt_tot_puppi --parallel 5
morph_parallel.py --output output --analysis mssm_vs_sm_heavy --eras 2016,2017,2018 --category_list ${CMSSW_BASE}/src/CombineHarvester/MSSMvsSMRun2Legacy/input/mssm_classic_categories.txt --variable mt_tot_puppi --parallel 5

mkdir -p output_mssm_vs_sm_heavy/combined/cmb/; rsync -av --progress output_mssm_vs_sm_heavy/201?/htt_*/*  output_mssm_vs_sm_heavy/combined/cmb/

Datacard creation for mssm_vs_sm_h125

morph_parallel.py --output output --analysis mssm_vs_sm_h125 --eras 2016,2017,2018 --category_list ${CMSSW_BASE}/src/CombineHarvester/MSSMvsSMRun2Legacy/input/control_region_categories.txt --variable mt_tot_puppi --parallel 5
morph_parallel.py --output output --analysis mssm_vs_sm_h125 --eras 2016,2017,2018 --category_list ${CMSSW_BASE}/src/CombineHarvester/MSSMvsSMRun2Legacy/input/mssm_new_categories.txt --variable mt_tot_puppi --parallel 5
morph_parallel.py --output output --analysis mssm_vs_sm_h125 --eras 2016,2017,2018 --category_list ${CMSSW_BASE}/src/CombineHarvester/MSSMvsSMRun2Legacy/input/sm_categories.txt --variable m_sv_puppi --parallel 5

mkdir -p output_mssm_vs_sm_h125/combined/cmb/; rsync -av --progress output_mssm_vs_sm_h125/201?/htt_*/*  output_mssm_vs_sm_h125/combined/cmb/

Datacard creation for mssm_vs_sm_CPV

morph_parallel.py --output output --analysis mssm_vs_sm_CPV --eras 2016,2017,2018 --category_list ${CMSSW_BASE}/src/CombineHarvester/MSSMvsSMRun2Legacy/input/control_region_categories.txt --variable mt_tot_puppi --parallel 5 --sm_gg_fractions ${CMSSW_BASE}/src/CombineHarvester/MSSMvsSMRun2Legacy/data/higgs_pt_v3.root
morph_parallel.py --output output --analysis mssm_vs_sm_CPV --eras 2016,2017,2018 --category_list ${CMSSW_BASE}/src/CombineHarvester/MSSMvsSMRun2Legacy/input/mssm_classic_categories.txt --variable mt_tot_puppi --parallel 5 --sm_gg_fractions ${CMSSW_BASE}/src/CombineHarvester/MSSMvsSMRun2Legacy/data/higgs_pt_v3.root

mkdir -p output_mssm_vs_sm_CPV/combined/cmb/; rsync -av --progress output_mssm_vs_sm_CPV/201?/htt_*/*  output_mssm_vs_sm_CPV/combined/cmb/

Workspace creation (exemplary for mssm_vs_sm_classic_h125)

ulimit -s unlimited
combineTool.py -M T2W -o ws_mh125.root  -P CombineHarvester.MSSMvsSMRun2Legacy.MSSMvsSM:MSSMvsSM --PO filePrefix=${CMSSW_BASE}/src/CombineHarvester/MSSMvsSMRun2Legacy/data/ --PO modelFile=13,Run2017,mh125_13.root --PO MSSM-NLO-Workspace=${CMSSW_BASE}/src/CombineHarvester/MSSMvsSMRun2Legacy/data/higgs_pt_v3_mssm_mode.root -i output_mssm_vs_sm_classic_h125/combined/cmb/ --PO minTemplateMass=110.0 --PO maxTemplateMass=3200.0

NB : for the mssm_vs_sm_CPV analysis, use mh1125_CPV_13.root instead of mh125_13.root and --PO MSSM-NLO-Workspace=${CMSSW_BASE}/src/CombineHarvester/MSSMvsSMRun2Legacy/data/higgs_pt_v3.root instead of --PO MSSM-NLO-Workspace=${CMSSW_BASE}/src/CombineHarvester/MSSMvsSMRun2Legacy/data/higgs_pt_v3_mssm_mode.root. The model building file to use is CombineHarvester.MSSMvsSMRun2Legacy.MSSMvsSM_CPV with the model named MSSMvsSM_CPV. You may also change ws_mh125.root to ws_mH1125_CPV.root. The command to run would then be

combineTool.py -M T2W -o ws_mH1125_CPV.root  -P CombineHarvester.MSSMvsSMRun2Legacy.MSSMvsSM_CPV:MSSMvsSM_CPV --PO filePrefix=${CMSSW_BASE}/src/CombineHarvester/MSSMvsSMRun2Legacy/data/ --PO modelFile=13,Run2017,mh1125_CPV_13.root --PO MSSM-NLO-Workspace=${CMSSW_BASE}/src/CombineHarvester/MSSMvsSMRun2Legacy/data/higgs_pt_v3.root -i output_mssm_vs_sm_CPV/combined/cmb/ --PO minTemplateMass=110.0 --PO maxTemplateMass=3200.0

Model-dependent CLs 95% limits for mssm_vs_sm_classic_h125

Computing limits:

mkdir -p calculation_mh125_mssm_vs_sm_classic_h125
cd calculation_mh125_mssm_vs_sm_classic_h125

ulimit -s unlimited
combineTool.py -M AsymptoticGrid ${CMSSW_BASE}/src/CombineHarvester/MSSMvsSMRun2Legacy/input/mssm_asymptotic_grid_mh125.json -d ${CMSSW_BASE}/src/output_mssm_vs_sm_classic_h125/combined/cmb/ws_mh125.root --job-mode 'condor' --task-name 'mssm_mh125_mssm_vs_sm_classic_h125_1' --dry-run --redefineSignalPOI x --setParameters r=1 --freezeParameters r -v1 --cminDefaultMinimizerStrategy 0 --X-rtd MINIMIZER_analytic --cminDefaultMinimizerTolerance 0.01

# After adaption of each shell script and condor configuration matching mattern condor_mssm_mh125_mssm_vs_sm_classic_h125_1.{sh,sub}, submit to batch system:
condor_submit  condor_mssm_vs_sm_classic_h125_1.sub

Collecting limits:

Basically the same command as above, but with a different task name (in case some of the jobs have failed)

ulimit -s unlimited
combineTool.py -M AsymptoticGrid ${CMSSW_BASE}/src/CombineHarvester/MSSMvsSMRun2Legacy/input/mssm_asymptotic_grid_mh125.json -d ${CMSSW_BASE}/src/output_mssm_vs_sm_classic_h125/combined/cmb/ws_mh125.root --job-mode 'condor' --task-name 'mssm_mh125_mssm_vs_sm_classic_h125_2' --dry-run --redefineSignalPOI x --setParameters r=1 --freezeParameters r -v1 --cminDefaultMinimizerStrategy 0 --X-rtd MINIMIZER_analytic --cminDefaultMinimizerTolerance 0.01

Plotting limits:

plotLimitGrid.py asymptotic_grid.root --scenario-label="M_{h}^{125} scenario (h,H,A#rightarrow#tau#tau)" --output mssm_mh125_mssm_vs_sm_classic_h125  --title-right="137 fb^{-1} (13 TeV)" --cms-sub="Own Work" --contours="exp-2,exp-1,exp0,exp+1,exp+2,obs" --model_file=${CMSSW_BASE}/src/CombineHarvester/MSSMvsSMRun2Legacy/data/mh125_13.root --y-range 2.0,60.0 --x-title "m_{A} [GeV]"

Model-dependent CLs 95% limits for mssm_vs_sm_h125

Computing limits:

mkdir -p calculation_mh125_mssm_vs_sm_h125
cd calculation_mh125_mssm_vs_sm_h125

ulimit -s unlimited
combineTool.py -M AsymptoticGrid ${CMSSW_BASE}/src/CombineHarvester/MSSMvsSMRun2Legacy/input/mssm_asymptotic_grid_mh125.json -d ${CMSSW_BASE}/src/output_mssm_vs_sm_h125/combined/cmb/ws_mh125.root --job-mode 'condor' --task-name 'mssm_mh125_mssm_vs_sm_h125_1' --dry-run --redefineSignalPOI x --setParameters r=1 --freezeParameters r -v1 --cminDefaultMinimizerStrategy 0 --X-rtd MINIMIZER_analytic --cminDefaultMinimizerTolerance 0.01

# After adaption of each shell script and condor configuration matching mattern condor_mssm_mh125_mssm_vs_sm_h125_1.{sh,sub}, submit to batch system:
condor_submit  condor_mssm_vs_sm_h125_1.sub

Collecting limits:

Basically the same command as above, but with a different task name (in case some of the jobs have failed)

ulimit -s unlimited
combineTool.py -M AsymptoticGrid ${CMSSW_BASE}/src/CombineHarvester/MSSMvsSMRun2Legacy/input/mssm_asymptotic_grid_mh125.json -d ${CMSSW_BASE}/src/output_mssm_vs_sm_h125/combined/cmb/ws_mh125.root --job-mode 'condor' --task-name 'mssm_mh125_mssm_vs_sm_h125_2' --dry-run --redefineSignalPOI x --setParameters r=1 --freezeParameters r -v1 --cminDefaultMinimizerStrategy 0 --X-rtd MINIMIZER_analytic --cminDefaultMinimizerTolerance 0.01

Plotting limits:

plotLimitGrid.py asymptotic_grid.root --scenario-label="M_{h}^{125} scenario (h,H,A#rightarrow#tau#tau)" --output mssm_mh125_mssm_vs_sm_h125  --title-right="137 fb^{-1} (13 TeV)" --cms-sub="Own Work" --contours="exp-2,exp-1,exp0,exp+1,exp+2,obs" --model_file=${CMSSW_BASE}/src/CombineHarvester/MSSMvsSMRun2Legacy/data/mh125_13_fixed.root --y-range 2.0,60.0 --x-title "m_{A} [GeV]"

Model-dependent CLs 95% limits for mssm_vs_sm_heavy

Computing limits:

mkdir -p calculation_mh125_mssm_vs_sm_heavy
cd calculation_mh125_mssm_vs_sm_heavy

ulimit -s unlimited
combineTool.py -M AsymptoticGrid ${CMSSW_BASE}/src/CombineHarvester/MSSMvsSMRun2Legacy/input/mssm_asymptotic_grid_mh125.json -d ${CMSSW_BASE}/src/output_mssm_vs_sm_heavy/combined/cmb/ws_mh125.root --job-mode 'condor' --task-name 'mssm_mh125_mssm_vs_sm_heavy_1' --dry-run -v1 --cminDefaultMinimizerStrategy 0 --X-rtd MINIMIZER_analytic --cminDefaultMinimizerTolerance 0.01

# After adaption of each shell script and condor configuration matching mattern condor_mssm_mh125_mssm_vs_sm_heavy_1.{sh,sub}, submit to batch system:
condor_submit  condor_mssm_vs_sm_heavy_1.sub

Collecting limits:

Basically the same command as above, but with a different task name (in case some of the jobs have failed)

ulimit -s unlimited
combineTool.py -M AsymptoticGrid ${CMSSW_BASE}/src/CombineHarvester/MSSMvsSMRun2Legacy/input/mssm_asymptotic_grid_mh125.json -d ${CMSSW_BASE}/src/output_mssm_vs_sm_heavy/combined/cmb/ws_mh125.root --job-mode 'condor' --task-name 'mssm_mh125_mssm_vs_sm_heavy_2' --dry-run --freezeParameters r -v1 --cminDefaultMinimizerStrategy 0 --X-rtd MINIMIZER_analytic --cminDefaultMinimizerTolerance 0.01

Plotting limits:

plotLimitGrid.py asymptotic_grid.root --scenario-label="M_{h}^{125} scenario (H,A#rightarrow#tau#tau)" --output mssm_mh125_mssm_vs_sm_heavy  --title-right="137 fb^{-1} (13 TeV)" --cms-sub="Own Work" --contours="exp-2,exp-1,exp0,exp+1,exp+2,obs" --model_file=${CMSSW_BASE}/src/CombineHarvester/MSSMvsSMRun2Legacy/data/mh125_13_fixed.root --y-range 2.0,60.0 --x-title "m_{A} [GeV]"

Model-dependent CLs 95% limits for mssm_vs_sm_CPV

Computing limits:

mkdir -p calculation_mh125_mssm_vs_sm_CPV
cd calculation_mh125_mssm_vs_sm_CPV

ulimit -s unlimited
combineTool.py -M AsymptoticGrid ${CMSSW_BASE}/src/CombineHarvester/MSSMvsSMRun2Legacy/input/mssm_asymptotic_grid_CPV.json -d ${CMSSW_BASE}/src/output_mssm_vs_sm_CPV/combined/cmb/ws_mH1125_CPV.root --job-mode 'condor' --task-name 'mssm_mh125_mssm_vs_sm_CPV_1' --dry-run -v1 --cminDefaultMinimizerStrategy 0 --X-rtd MINIMIZER_analytic --cminDefaultMinimizerTolerance 0.01 --redefineSignalPOI x --setParameters r=1 --freezeParameters r

# After adaption of each shell script and condor configuration matching mattern condor_mssm_mh125_mssm_vs_sm_CPV_1.{sh,sub}, submit to batch system:
condor_submit  condor_mssm_mh125_mssm_vs_sm_CPV_1.sub

Collecting limits:

Basically the same command as above, but with a different task name (in case some of the jobs have failed)

ulimit -s unlimited
combineTool.py -M AsymptoticGrid ${CMSSW_BASE}/src/CombineHarvester/MSSMvsSMRun2Legacy/input/mssm_asymptotic_grid_CPV.json -d ${CMSSW_BASE}/src/output_mssm_vs_sm_CPV/combined/cmb/ws_mH1125_CPV.root --job-mode 'condor' --task-name 'mssm_mh125_mssm_vs_sm_CPV_2' --dry-run -v1 --cminDefaultMinimizerStrategy 0 --X-rtd MINIMIZER_analytic --cminDefaultMinimizerTolerance 0.01 --redefineSignalPOI x --setParameters r=1 --freezeParameters r

Plotting limits:

plotLimitGrid.py asymptotic_grid.root --scenario-label="M_{H_{1}}^{125} CPV scenario (H_{1},H_{2},H_{3}#rightarrow#tau#tau)" --output mssm_mh125_mssm_vs_sm_CPV  --title-right="137 fb^{-1} (13 TeV)" --cms-sub="Own Work" --contours="exp-2,exp-1,exp0,exp+1,exp+2,obs" --model_file=${CMSSW_BASE}/src/CombineHarvester/MSSMvsSMRun2Legacy/data/mh1125_CPV_13.root --y-range 1.0,20.0 --x-title "m_{H^{#pm}} [GeV]" --mass_histogram m_H1

Adaptions needed for job submission via htcondor

The following line in file CombineToolBase.py from the CombineHarvester package needs to be updated to be able to run on htcondor batch systems with the following configuration snippet:

ETP batch system

Requirements = ( (Target.ProvidesCPU == True) && (TARGET.ProvidesEKPResources == True ) )
universe = docker
docker_image = mschnepf/slc7-condocker
+RequestWalltime = 10800
+ExperimentalJob = True
RequestMemory = 2000
RequestCpus = 1
accounting_group = cms.higgs

No additions needed for CERN or NAF batch systems

Analysis of predictions for SM-like Higgs boson

There is a script to analyze the predictions for the SM-like Higgs boson from the various MSSM benchmark scenarios. It can be used to plot the total scale factor to be applied to the SM(125.09) templates for ggphi and qqphi, as well as BSM contributions to its individual parts (without mass effects), and mass effects going into the scale factor by using SM(125.09) predictions as reference.

The script is called plotSMlikeHiggsPredictions.py. To use it, please open a fresh terminal to be able to source and use latest ROOT:

source /cvmfs/sft-nightlies.cern.ch/lcg/views/dev3/latest/x86_64-centos7-gcc10-opt/setup.sh

Then, to get the full ensemble of plots, use the following set of commands executed from your current ${CMSSW_BASE}/src folder. Please note, that ${CMSSW_BASE} is not set (you should not execute cmsenv), it serves rather as placeholder for the absolute path of your analysis setup.

# models with the light scalar h being SM-like
for m in CombineHarvester/MSSMvsSMRun2Legacy/data/mh125*.root;
do
    python CombineHarvester/MSSMvsSMRun2Legacy/scripts/plotSMlikeHiggsPredictions.py \
        --mssm-benchmark "${m}" --bsm-sm-like h \
        --sm-predictions CombineHarvester/MSSMvsSMRun2Legacy/input/sm_predictions_13TeV.json; 
done;

# models with the heavy scalar H being SM-like
for m in CombineHarvester/MSSMvsSMRun2Legacy/data/mHH125*.root;
do
    python CombineHarvester/MSSMvsSMRun2Legacy/scripts/plotSMlikeHiggsPredictions.py \
        --mssm-benchmark "${m}" --bsm-sm-like H \
        --sm-predictions CombineHarvester/MSSMvsSMRun2Legacy/input/sm_predictions_13TeV.json;
done;

# CP violating models with H1 being SM-like
for m in CombineHarvester/MSSMvsSMRun2Legacy/data/mh1125*.root;
do
    python CombineHarvester/MSSMvsSMRun2Legacy/scripts/plotSMlikeHiggsPredictions.py
        --mssm-benchmark "${m}" --bsm-sm-like H1
        --sm-predictions CombineHarvester/MSSMvsSMRun2Legacy/input/sm_predictions_13TeV.json;
done;