Another simple algorithm would be to calculate the distance between the highest energy strip and the mean of the distribution (from here I'll refer to this as MMM for "mode minus mean"),

The idea here is that in a pion shower the mean would be between the two decay photon showers while the highest energy strip will be in one of the decay photon showers, leading to large values of MMM.  In a single direct photon shower, however, MMM should vanish.
    It should be clear, however, that this variable is extremely vulnerable to bad SMD strips.  If the high strip in a direct photon shower is lost due to poor status then the calculated MMM will be artificially high.   
    Despite this potential flaw the preliminary results aren't entirely terrible.  The analysis was performed on the same simulation files as in the previous few studies.