MC BDT numbers in third and fourth preshower bins

(I) Variables used for training

(1) nCh (2) nBtow (3) nEtow (4) ptBalance (5) R3x3jet7 (6) R2x2jet7 (7) R2x1jet7 (8) R2x1c3x3 (9) R2x1c2x2
(10) R2x2c3x3 (11) nU (12) nV (13) fiveStripU/twentyfiveStripU (14) fiveStripV/twentyfiveStripV (15) ePre1
(16) ePre2 (17) ePost

(II) Preselection Cuts

(i) z vertex cut -100cm<z<100cm

(ii) Neutral fraction of jet matched to photon >0.9

(iii) cos(phi_gamma - phi_jet)<-0.8

(iv) |eta_jet|<0.8

(v) ptGamma > 7.0 GeV

(vi) ptJet > 5.0 GeV

(vii) 1.08 < eta_gamma<2.0

(viii) Number of hits in each SMD plane > 3

(ix) Away side jet neutral fraction <0.9

I) Background rejection vs signal efficiency plots
 

Third preshower bin                                                                                  Fourth preshower bin
 

II) BDT discriminant output
Third preshower bin                                                                                  Fourth preshower bin
 

III) Efficiency plots
Third preshower bin                                                                                  Fourth preshower bin
 

Note: Numbers quoted in this table are what we hope to get after scaling prompt photon and QCD background to 3.1 pb-1.
We still have huge QCD background and should probably think of running private simulation with parameters adjusted to pytune 320(perugia0).