Just to let you know I started playing around with the available Klong selector from Antimo (as last resort for our B0 puzzle). The algorithm returns a likelihood according to EMC quantities and a likelihood according to the IFR quantities associated with it. You can see the distribution for the EMC (left column) and the IFR (right column) and for the clusters truth-matched as photons (top row) or as KL or neutron (bottom) http://www.slac.stanford.edu/~rfaccini/phys/vub/xmas02/ifremclike.eps While the IFR one might have a problem, the emc selector seems promising. The correlations with the variables we are using so far in selection are shown in http://www.slac.stanford.edu/~rfaccini/phys/vub/xmas02/corrlike.eps I will start playing aroung with replacing the existing shape cuts with this emc likelihood comments and andvice are welcome ciao ric