The MIP peak calibration fit can be used to estimate the efficiency of veto settings for the ACD. The fit is a falling exponential plus a gaussian. The Gaussian represents the charged particles which we want the veto to reject. The falling exponential has three components: backsplah, residual signal from previous particles, and backsplash. In general we do not want to reject these components, since not falsely rejecting backsplash is one of the objectives of the ACD. 
The efficiency plot below is for one set of PMT's on the top panel.