Multiple event counting

[mizorafa]

issue:
In the stacked & exclusion-map mode, the current scheme first makes a run-wise map by always stacking neighbouring runs to the target run, and then stacks all the neighbours-stacked maps among runs. This leads to a situation where some of events/pixels especially outside the exclusion masks are used/counted multiple times in the final outcome. This situation may potentially bring about the following problems:

• Runs that have more acceptable neighbours are used more. This may create a bias to the bkg model.
• More importantly, the stacked counts that will be used to compute the rate for the final output model will not obey the Poissonian. this may be not easy for most users, as the current analysis after DL3 in LST presumes that the bkg model has a statistical uncertainty following the Possion distribution.