interested people: Emanuele, Luca, Ben, Efe, Tomas Jezo, Markus Seidel, Alexander Grohsjean, Ludovic Scyboz, Philippe G., ADD YOUR NAME HERE

A discussion is scheduled for Monday morning.

The more recent results from the ongoing (TH) study using the powheg-box-res framework (developed in https://arxiv.org/abs/1509.09071 and https://arxiv.org/abs/1607.04538) can be found here (P. Nason talk) https://indico.cern.ch/event/596233/timetable/

Possible topics for LH:

• experimental needs

• availability of benchmarks results

• dedicated interfaces for NLOPS matching (availability, missing/desired features). We had a discussion on LHE v3 on Saturday, which has also to do with this point…

• uncertainties

• comparison powheg vs mc@nlo (any news from herwig, or sherpa?)

Discussion on Monday:

overview: slides

ATLAS perspective: slides

technicalities (bb4l generator):

• make sure that numerical accuracy reached in event generation from MC is the same as in TH-paper

• possible to have grids from the authors (seems to be a viable solution, as atlas and cms will agree - or already agreed - on the settings)

• if grids will be provided, need to agree on parameters to scan upon. It seems that order 20 runs will be enough (a scan would mostly be done using 5-10 values for mtop, and 2-3 values for hdamp)

• It might even be possible to use the powheg reweighting machinery to avoid having to re-run all the grids. This might depend on how far mtop is moved from the central value. T. Jezo and collaborators have tried this. Perhaps it'd be useful to perform a closure-test, but using the outermost mtop values that atlas/cms would use.

• this reweighting would miss the mtop dependence in R/B. There exists an experimental facility in powheg, to capture these effects via a reweighting, but it was implemented only for DY, and rarely used. Not clear it would work here.

[ER: to be continued]