Shower splitting function enhancement in the context of ME+PS merging

[Frank Siegert]

Currently the enhancement of splitting functions in the shower (e.g. for photon production) is not designed to be used with ME+PS merging. The problem is in the Sudakov rejection of ME-regime emissions which can't be simply applied to enhanced emissions. The emission will have to be completely unweighted in case it ends up in the ME regime (and thus will be rejected).

[Stefan Hoeche]

Sudakov rejection arising from a jet veto in the shower must finally produce P_ME in Eq.(21) of arXiv:0903.1219. For weighted events from the shower this can either be achieved by using a non-uniform rejection weight, or by partial unweighting. The first method would imply a bookkeeping of weights associated with vetoed events. This results in reweighting all events from a given ME-level process (i.e. the prescription used in Sherpa v1.0.x, which was unfavourable because it is very difficult to deal with on the analysis side). The second method would mean to unweight all shower emission _trials_ which would fall into the ME domain, no matter whether they finally pass the f/g rejection in Eq.(20) of arXiv:0912.3501. Therefore one _must_ construct the emission, even if it is rejected in the f/g step, to decide whether it belongs to the PS or the ME domain (and whether the g/h weight is to be applied analytically or via MC rejection, respectively). This is cumbersome and presumably slows the shower down significantly.

[Stefan Hoeche]

This won't happen, and we have moved away from using SF enhancement anyway.