LHC EWWG aTGC ATLAS+CMS+TH discussion

Wednesday 13 Jan 2016, 10:00 → 18:00 Europe/Zurich

4/3-004 - TH Discussion Room (CERN)

Present:
Senka Djuric,  Adam Falkoswski, Lindsey Grey, Vasiliki Kouskoura, Josh Kunkle, Kirstin Lohwasser, Alexander Oh, Francesco Riva

Discussion:
Limit setting
How to mitigate the overconstraining of anomalous couplings in case of downward fluctuations of data?
AI: Should revisit CLs limits for next meeting.
Add a section on limit setting for aGC parameter in document.

Scale cut-off
The EFT range of validity is varying with the scale Lambda. To exploit the full sensitivity range to new physics the limits should be given as a function of the cut-off scale. Two possibilities to probe the scale are suggested:
a) cut-off events in data with sqrt(s)>Lambda. Pro: cleanly restricted energy range, con: difficult to implement experimentally for final states with neutrinos.
b) introduce a form-factor to set the contribution of new physics operators to 0 for sqrt(s)>Lambda. Pro: no need to cut on data, con: leakage into higher sqrt(s) region.
Needs a toy study to see what performance to expect from a) and b).
AI: Make a toy study of a) and b).

Generators
Matrix element implementation for Higgs and aTGC dim6 basis available in Madgraph. Not available yet is the calculation of weights. Would be useful to have the event-by-event reweighting possibility as in McAtNlo available also in Madgraph.
AI: Follow up how reweighting can be implemented in Madgraph.

Basis are equivalent if complete. Practical problem arises that not all operators can be constrained in single channel analysis (e.g. higgs operators in WZ final state). Ideally a global fit to all channels would provide the best constraints. For practical purposes need to identify operators to which a given final state is sensitive.
Q: What is a sufficient set of information to be able to combine limits between SM and Higgs analysis?

For the extraction a description of the differential cross section dependence on the couplings is needed. Two approaches are available:

a) Morphing, i.e. a spline interpolation between discrete coupling values.
b) Event reweighting, i.e. storing enough information for each event to be able to reweight to any coupling value.
Agreed that both approaches work and are valid.

 

GIT repo for recommendation Document

https://gitlab.cern.ch/jkunkle/LHEEWKWorkingGroup.git

 

 

 

 

 

 

10:00 → 11:30 Roundtable