The events are read once to construct the run summary and then one more time each time the run is selected as a neighbor to another run, so worst case for each other run / n^2 times.
This slows things down significantly (citation needed, but at least for LST dl3 I would be very surprised if this is not the main reason).
The events are read once to construct the run summary and then one more time each time the run is selected as a neighbor to another run, so worst case for each other run /
n^2times.This slows things down significantly (citation needed, but at least for LST dl3 I would be very surprised if this is not the main reason).