Just like the last time his happened, the left arm ROCs are in sync and the right arm ROCs are in sync, so we don't have to worry about mixing up data from different detectors. But if we're getting extra events in the buffers, then we may not know which trigger type is associated with the data in the spectrometer. Since we had large prescale factors (up to 70), we could have a problem knowing how to scale the different event types.
Since the high rate is in the left arm which has very few pions, we decided to start running with all prescale factors set to one. This only increases the rate slightly, since most of the prescaled T7 events are also good T3 events. It yields an apparent deadtime of near 50% (because most events appear as T3 AND T7, but only get read out once). We'll check that the good event fraction stays at ~8% to verify that the large apparent dead time isn't real, and then stay this way for this setting.