ATLAS Contribution

• Numbers shows on slide 1 are from H->γγ only, extrapolated from the Moriond CONF note

CMS contribution

• Results are from the ttH(bb) semileptonic mode only; the DNN is not yet optimized to reduce confusion in top/Higgs reconstruction.
• slide 5: The table shows how each truth bins distributes into each reco bin (each line sums to 1). The non-diagonal terms are much larger here than in the 3-bin case.
• The 4-bin analysis may be difficult for ttH(bb). However DNN could possibly be improved (ATLAS expects better performance - it would be desirable to have explicit studies/confirmation)

Discussion

• ttH binning:
  • Decision to go for Option 1 since smaller bins are safer with respect to future evolution (in particular 200-300 GeV instead of 200-350 GeV). Could merge 200-300 GeV and 300-400 GeV bins if needed.
• ggF binning:
  • the binning decisions mainly concern the 400-500 GeV region, which is at the transition between the H->γγ/ZZ and H->bb sensitivity regions: H->γγ/ZZ lose sensitivity above 400 GeV while H->bb sensitivity starts around 450 GeV (see slides 3 and 5 in https://indico.cern.ch/event/827970/contributions/3465224/attachments/1860897/3058317/2019-06-12_ggH_LH.pdf).
  • However H->γγ/ZZ is only measured in the 350-600 GeV bin so not clear where the sensitivity stops exactly, and similarly H->bb sensitivity below 450 GeV is not known. Would need more information on sensitivities in this region to understand which solution entails the least extrapolation.
  • ATLAS could have updated H->bb sensitivity estimates soon.
  • Option 1 seems preferable (smaller bins), but this can be revisited depending on the results of the further studies.
• ggF jet bins:
  • pTJ cut could be relative to the center of the pTH bin, or the event-by-event pTH; should not have much impact since the pTH bins are rather narrow
  • H->WW should be the analysis most affected, some studies would be useful
  • Could consider raising the pTJ cut even in the 120-200 GeV bin

Action items:

• Would be good to have H->WW studies from ATLAS and CMS: ideally per-pTH optimizations of the pTJ cut; or more simply r distributions for Higgs signal and backgrounds (mainly ttbar) in coarse bins of pTH.
• also additional H(bb) studies from ATLAS ?

Aim to converge on the gg->H binning at the next meeting, planned for end of August. Will also have a discussion of the variables to report in the diffXS measurements, following up from LH (Pedja will provide the picture of the blackboard in LH...)