Effect of PV reco inefficiencies on fake SVs (and its monitoring)
Follow up of discussion started here: slides.
General description: The type and configuration of PV reco algorithms seem to have a large effect on the occurrence of fake SVs. Current PV testing and monitoring does not fully seem to catch PV reconstruction inefficiencies that lead to fake SVs that have the potential to greatly affect ghost rate / selection / BW division / etc...
Should a dedicated test be made? Maybe in PRDashboard? Or can existing tests be adapted to better catch this?
fyi: @gciezare @msaur @mgiza @adudziak @dovombru @mvesteri @poluekt (feel free to adapt)