====== gg -> ZH + 0/1 Jet ======

**Project Description**

Look at MC multileg 0+1jet@LO description of ggZH process, compare nominal prediction vs simple LO MC.
From QCD scale variations derive estimate of uncertainties for STXS VH bins and quantify the improvement from LO MC to 0+1jet@LO multileg MC.

**Interested Participants**

Carlo Pandini, Emanuele Re, Milene Calvetti, Valerio Dao, Enrico Bothmann, Paolo Francavilla

**Preliminary Sherpa results**

Here is a first proof-of-principle nominal prediction for 0+1jet@LO with Sherpa:
  * [[https://www.theorie.physik.uni-goettingen.de/~bothmann/les_houches_2019_zh_gluon_initiated/rivet-plots/|Rivet Plots]]
  * [[https://www.theorie.physik.uni-goettingen.de/~bothmann/les_houches_2019_zh_gluon_initiated/setups/zh/zh.dat|Config File]]

**Points to consider**
  * check that e.g. powheg+py8 and mg5+py8 give the same, if the shower starting scale is the same
  * 2 "substudies": 1 targeted to just the [[https://twiki.cern.ch/twiki/pub/LHCPhysics/LHCHXSWGFiducialAndSTXS/simplifiedXS_VH_1.1.pdf|STXS bins]], one with some more differential distributions
  * tools: powheg+py8 (current baseline for ATLAS/CMS), mg5+py8 (0 jet / 0+1 jet LO merging), sherpa (0 jet / 0+1 jet LO merging)
  * for this study: H undecayed, Z decaying to 1 lepton family. Check that stxs rivet routine can do this (for py8 as well as sherpa). Otherwise, can also do H->bb, but in this case we need to have the complete information in the HepMC also for sherpa, as Rivet looks for Higgs and Z as the last entry with correct id before the decay..
  * for STXS unc.: consider jet pT spectrum (for definition of nJet bins) and impact on uncertainties
  * for mg5+py8 validation: m(VH) is a good variable to check inclusive / 0+1jet multileg agreement

**Rivet code**
  * [[https://gitlab.cern.ch/LHCHIGGSXS/LHCHXSWG2/STXS/Classification|STXS-rivet routine]]: keep unchanged to run over STXS bin selection
  * VH-HXSWG-rivet routine (will add link here): use this to include additional variables, and any modification we might want to consider
  * we can for each MC setup run both routines, to get results for both "substudies"