Simplified models in the SUSY context

Interested people: Suchita, Wolfgang, Sabine, Aoife, Sofio, Tobias …. (please add your name!)

SMS wishlist talk

SUSY simplified model results are a generic and powerful way to constrain MSSM parameter space. Various simplified models topologies are considered by both ATLAS and CMS collaborations. These topologies often include more than two free parameters eg. the mass of intermediate particle(s) in addition to the mass of the mother and LSP in case of cascade decays. It is possible to Interpolate between the 95% CL upper limit results for different mass(es) of intermediate particle(s) in order to obtain generic result for any intermediate masse(es). In this way, the SMS results can be applied to more generic

The following wish list is based on this principle and aims towards making the simplified models results more accessible and useful. For each of the points on the wish list, we list one potential candidate where the existing SMS results can be improved.

1. Digitize… digitize… digitize….
   Please provide the information contained in the cross section limit plots in electronic form, either as simple ascii file or a Root file. This is increasingly done by CMS, a good example is SUS13011, ATLAS is still dragging behind. (this is gradually improving…)
   
   
2. For topologies involving cascade decays, provide results for more than one (at least 3) intermediate mass values.
   Results on processes involving 2-step decays (e.g. trileptons from EWino decays via sleptons) crucially depend on the intermediate mass value, because this determines the kinematics of the events. If results are given for only one mass fraction value, we cannot use them, because we cannot interpolate. This is for example the case for the 3taus+MET results from CMS, SUS12022, which are available for x=0.5 only. Such results which have only one value of x are in fact completely useless for our purpose. In order to allow interpolation for different intermediate mass values, least three mass fraction are needed. We recommend to provide the results for x = 0.05, 0.5, 0.95 (or if possibe x = 0.05, 0.25, 0.5, 0.75, 0.95).
   
   
3. Provide good coverage of the parameter space considered
   Large changes in the excluded cross sections (or holes in the parameter space), as present e.g. in this example from the EWino search in ATLAS, cannot be reliably interpolated. Please provide uniform bins or, if this is not feasible, ensure a good coverage of the sensitive region.
   
   
4. Avoid too restrictive assumptions
   When presenting an SMS interpretation result, the underlying physics considered should be as generic as possible. Instead of fixing certain aspects of the scenario, it would be more helpful to discuss “default” and extreme cases. For example:
   • Same-sign leptons plus jets from gluino and squark production: in Fig. 16 of ATLAS-CONF-2013-007, two mass parameters are fixed (see point 2. above), and flavor-democratic slepton decays are assumed –> Suchita's nightmare analysis.
   • On the other hand, the CMS search for electroweak production of charginos, neutralinos and sleptons using leptonic final states, SUS12022, discusses several extreme cases for slepton decays, which is a more helpful approach.
     
     
5. For topologies involving different decays on each leg, parametrize results in terms of branching fractions.
   A typical example is stop1 pair production with one stop decaying into top+neutralino1 and the other one into bottom+chargino1. The final state is tb+MET, but the constraints on the cross section depend on 3 masses (stop1, chargino1, neutralino1) as well as the 2 branching ratios. If only the two decay channels are open, BR(stop1 –> bottom+chargino1) = 1-BR(stop1 –> top+neutralino1), leaving us with 4 free parameters. It would be useful to present results in, e.g., the stop1 versus neutralino1 mass plane for different values of branching ratios.
   
   
6. Give expected upper limits on sigma x BR in addition to the observed ones
   The expected upper limits are needed in order to identify the most sensitive topology. This is crucial for a correct statistical treatment of complex spectrum decompositions, which may be constrained by more than topology.
   
   
7. Incorporate likelihoods
   An approximate form of the final likelihood (e.g. through RooFit/RooStats) or at least more than one CL UL on the cross sections would be extremely welcome. This would help e.g. assessing which topologies should be combined.
   

Finally we note that efficiency maps would be very useful for enlarging the scope of SMS interpretations, cf CERN workshop.