Looks like the full event list was being used instead of the mask. The tracking and muonID are inside the mask, so should be the correlated chi2 algorithm.
@raaij @acasaisv