====== Discussion on Pseudo Observables/Simplified Cross Sections for Higgs Measurements ====== //If you are interested in contributing please subscribe to the [[https://e-groups.cern.ch/e-groups/EgroupsSearch.do?searchValue=houches-2015-topics-higgs-simplified-xsection|mailing list]]. You will be asked to log in with your CERN account. If you don't have a CERN account, please fill [[https://account.cern.ch/account/Externals/RegisterAccount.aspx|this form]] to create a lightweight account.// //This twiki page is closed -- session 2, we welcome your comments, and in particular suggestions on the BSM-motivated regions including proposals for the "BSM jokers" [[2015:groups:pseudoxsecsBSM|here]].// //[Update]//: For an overview talk with some more details see [[https://indico.cern.ch/event/399923/session/3/contribution/20/material/slides/0.pdf|here]] (corresponding to 2nd iteration below). ==== Goal: ==== Define "simplified cross sections" as general framework for Higgs measurements and combination of decay channels. * Provide "measurement interface" between raw experimental categories and final interpretation. * Can think of these as "differential/fiducial mu" per production channel, but normalized as cross sections (i.e., without dividing by SM predictions) * Ensure long shelf life of experimental measurements * Minimize theory dependence (theory systematics and model dependence) in measurements. * Can then be interpreted in SM (analogous to current mu fits) or different BSM scenarios (Higgs-EFTs or specific models) ==== Guiding principles for definition of simplified cross sections: ==== * Should be reasonably well constrained/measured * Important to think which phase-space regions are important to separate out from the theory side * where are largest theory systematics (e.g. ggF 0jet bin) * BSM sensitivity/interpretation (e.g. EFT breaks down at high energy). == Notes: == * Does not replace measurement of fiducial/differential cross sections in gammagamma and ZZ * Allows maximizing experimental sensitivity, e.g. can still use MVA-based selections, ... * Still important to ensure that MVAs do not introduce uncontrolled theory systematics (e.g. check which phase space regions get selected) * Some of the (pseudo)observables might also be * limits * asymmetries * continuous parameters for kinematic deviations (e.g. CP odd admixture) * Definition can evolve * Can split into more fine-grained bins once statistics allows (previous measurements remain useful) == Specific points for further discussion: == * Experiments will have to give correlations, need to define how to do the exact "information transfer" * in some cases covariance matrix may not be enough and need full likelihood? ==== 1st iteration ==== of possible cross sections (in the following "bins"), to be continued ... For 2nd iteration: Reduce number of bins, classify according to importance/feasibility == Notes: == * "ggF-topology", "VBF-topology", ... describe the SM-like kinematic topology, not necessarily only the SM production mode. So one can think of it as using the SM itself as a "simplified kinematic model". (In practice, on the experimental side the corresponding SM processes are used to evaluate and unfold the acceptance corrections.) * everything is meant schematically, all numbers are just for illustration * In the future, can add additional non-SM-like kinematic templates (e.g. CP-odd ggF) == ggF-topology == =0j \\ =1j, separate bin for pTj>~100 GeV (or pTH>~100 GeV) \\ =2j, separate bin with VBF cuts \\ >=3j \\ * inputs: gammagamma, ZZ, WW, tautau (for 1j > 100 GeV, 2j VBF and >=3j) == VBF-topology == main: mjj >~ 300 GeV (and have some rest 120-300 to not lose any region) \\ more possibilities: mjj, pTj, Deltaetajj \\ pTj1 = [0, 100, 200, 300, inf] \\ some jet-veto-like selection (e.g. pTHjj) \\ + kinematics measure for non-SM-like continuous shape changes (O(<5) parameters) * inputs: gammagamma, ZZ, WW, tautau, bbbar == VH-topology (V -> hadrons) == mjj 70 - 120 GeV \\ m(VH) = [0, 300, 600, inf] (more ideal from theory side) \\ or pT(V) = [0, 100, 200, inf] (matches analysis more closely) \\ and split everything by =0j, >=1j (at first only for high pT(V) or m(VH)) \\ * inputs: gammagamma, ZZ * Need to think about how to best deal with the overlap between VBF and VHhad == VH-topology (V -> leptons) == keep the Z and W separately: Z->ll, Z->nunu, W->lnu \\ same bins as for VHhad (need to see how well this works with neutrinos) \\ * inputs: bbbar, gamgam, ZZ, WW, tautau == gg->ZH-topology == as for VH, likely fewer bins (more relevant for higher pT(V), m(VH)) \\ == ttH == 2j, 3j, 4, 5j, 6j \\ + continuous CP-odd parameter \\ == bbH == single inclusive cross section bin == tHW, tHjq == inclusive and then as few kinematic parameters as possible ==== 2nd iteration ==== == Notes: == * jets are assumed as anti-kt 0.4, |eta|<4-5, pT>=25-30 GeV * bins are exclusive (in the selection, start from the lowest bin on the list) * "ggF-topology", "VBF-topology", ... describe the SM-like kinematic topology, not necessarily only the SM production mode. So one can think of it as using the SM itself as a "simplified kinematic model". (In practice, on the experimental side the corresponding SM processes are used to evaluate and unfold the acceptance corrections.) * explicit numbers are only meant as ball-park numbers Proposal for two reduced scenarios, "small" and "medium", trying to take into account feasibility with current statistics, theoretical uncertainties, and sensitivity to BSM. == ggF-topology == SMALL\\ 0j\\ >=1j, split into pTH < = mH, pTH = mH - 200 GeV, pTH > 200 GeV (highest bin for BSM)\\ >=2 VBF (mjj, Deltaetajj cuts as for VBF), split into =2 and >=3 by a pTHjj cut\\ MEDIUM\\ split pTH > 200 GeV into 200 GeV - high, and > high\\ split >=1j to =1j and >=2j (with the pTH splitting)\\ == VBF-topology == SMALL\\ rest\\ >=2j, with mjj and Deltaetajj cuts, split into =2 and >=3 by a pTHjj cut\\ "high-q^2" bin (can be replaced by something else BSMy, could be high pTj1)\\ MEDIUM\\ allow for continuous parameter to allow for contribution of CP odd coupling (interference is probably irrelevant, but could check)\\ another "BSM joker" (TBD by session 2)\\ == VH-topology == SMALL\\ exclusive in ll+nunu, lnu, had\\ pT(V) = 0 - mH (low) and mH - inf (high)\\ split high in =0j, >=1j\\ pT(V) > 200 GeV (for BSM) MEDIUM\\ split ll and nunu\\ split low pT(V) bin into 0j and >=1j\\ == gg->VH-topology == same as for VH, if completely degenerate can give the sum of VH and gg->VH == ttH == SMALL\\ 1 incl bin\\ MEDIUM\\ split into 0j, >=1j additional jets (with a rather high pT cut, maybe 50-100 GeV?)\\ add "BSM joker" (to be defined by session 2) == bbH == MEDIUM\\ 1 incl bin == tH == MEDIUM\\ 1 incl bin