There were a couple of errors in the referenced analysis, so below is the line of reasoning for calculating the trigger efficiency:
N3 = number of T3-only triggers = Ni*eS*eC*(PS-1)/PS N2 = number of T2-only triggers = Ni*eS*(1-eC)/PS N23= both T2 and T3 fired evts = Ni*eS*eC/PSwhere Ni is the "pristine" number of elastic events (needed from a calculation), eS is the single-arm (BB) trigger efficiency, eC is the coincidence trigger efficiency (prob. of detecting proton in the ND), and PS is the prescale factor for T2 events. We're assuming the prescale for T3's is 1. I'm also neglecting the livetime contribution, since it will simply scale up the efficiency.
In 167520 we found eC=26%. Given:
N3 = 9334 N2 = 2601 N23= 1043 and Ni = 73000 (from a calculation neglecting radiative effects and with rough geometrical acceptances).
Solutions for eS are:
eS = (N2 + N23)*PS/Ni -> eS = 0.499 eS = N3/(eC*Ni) * PS/(PS-1) -> eS = 0.526 eS = (N3 + N2 + N23)/Ni * 1./(eC*(PS-1)/PS + 1/PS) -> eS = 0.518
So, NOT correcting for livetime (~20%?) or radiative effects, we are
getting an electron-detection efficiency from run 2485 of ~50%.