Konigsberg used CDFSIM to try to reproduce rates in the CMX, and his numbers (and Melissa's) come close to the data. It was suggested that I do the same. I decided to compare my CMX rates with those Konigsberg predicted for Run II with no snout.
He required that hits arrive within 30 nsec of the crossing, and found that most of the hits came from the plug edge, beam pipe, and back-scatter from the toroids. The occupancy he predicted was 0.11 hits per minimum bias event.
I predicted a drift chamber occupancy of 0.15+-0.03, which is close, and found most of the prompt hits came from the beam pipe and back-scatter. The scintillator occupancy comes up as .03
However, when I simulate the effect of previous events by requiring instead of arrival time less than 30 nsec, arrival time MODULO 396 nsec be less than 30 nsec, the number of scintillator hits increases substantially, to 0.19 per minimum bias event. Most of these are due to neutron interactions in the scintillator, and won't, except by accidental coincidence, mimic a trigger.
We can reduce the rates with a tighter gate--demanding that hits appear after 10 nsec and before 30, for example.