Skip to content

Systematic uncertainties + electroweak corrections

Nicolas Postiau requested to merge npostiau/shears:systematic_uncertainties into master

First implementation of the electroweak corrections and of the framework to run on the systematic uncertainties:

  • Electroweak corrections are implemented for ZZ and WZ (both virtual and gamma-induced). This includes the 2 tables added (in "Data").
  • You can now run on the systematics interactively, or with the prepareAllJobs.py script. The .sh submission scripts have not been upgrated yet. For now, only the electroweak corrections are implemented.

In order to launch the analysis on all the systematics:

./Tools/prepareAllJobs.py --listDataset test_listSamples.txt --suffix aTest --syst all cd OUTPUTS/aTest big-submission sendJobs_aTest.cmd cd ../.. ./Tools/prepareAllJobs.py --listDataset test_listSamples.txt --suffix aTest --syst all --harvest

This creates files for all the systematics listed in the systList.txt file, for all processes. There are also files labeled as "final", which contain ALL the shapes for all systematics.

Notice that by default, only the final mT plot is kept when we run on ALL systematics (for everything produced in this submission). This can be changed very easily in the script (only 1 line). Let me know if you prefer to add an option.

If you want to run only on a few systematics, here is the syntax :

./Tools/prepareAllJobs.py --listDataset test_listSamples.txt --suffix aTest --syst ewk_up -> only on ewk_up

./Tools/prepareAllJobs.py --listDataset test_listSamples.txt --suffix aTest --syst ewk -> ewk_up and ewk_down

./Tools/prepareAllJobs.py --listDataset test_listSamples.txt --suffix aTest -> nominal (no systematics)

./Tools/prepareAllJobs.py --listDataset test_listSamples.txt --suffix aTest --syst no -> nominal (no systematics)

./Tools/prepareAllJobs.py --listDataset test_listSamples.txt --suffix aTest --syst all -> all the systs in the systList.txt file, plus nominal. In this case, only the final mT plots are kept.

The idea is that when we launch the analysis on "all", then we have all the final histograms that we need for the datacards.

PLEASE DON'T REVIEW YET. There are a lot of things to improve/develop/clean up and this MR will evolve in the following days.

Edited by Hugo Delannoy

Merge request reports