In order to help stabilize the latest fits, it is desirable to see if we can reduce the conversion and low mass backgrounds.

---

Changes in Cuts

General ideas:

• To reduce backgrounds of several types
  • Remove the (Energy in SMD u+v) / (Tower Energy) > 0.006 cut -- it doesn't seem to do much
  • Try requiring a sector to have no more than 2 u SMD clusters and and no more than 2 v SMD clusters
• To reduce conversions
  • Bring the preshower energy threshold to something lower (a few MeV)--currently the threshold is at 40 MeV
• To reduce low mass peak
  • Raise the SMD seed energy from 2 MeV to something higher (such as 5 MeV)

---

Cut options considered

Cut Set Name	(SMD u+v)/E	# Clus per Layer	Preshower E [GeV]	SMD Seed E [MeV]
old	0.006	---	40	2
new 1	---	---	5	5
new 2	---	---	4	5
new 3	---	2	3	5
new 4	---	---	40	2
new 5	---	---	3	2
new 6	---	2	40	2
new 7	---	2	3	2

---

Results

The PDF files are arranged as in this blog, and are attached at the bottom.  Note: the new 1 & 2 had the initial histograms generated in December using the L0 level 'h' for the MC portion, while the other results all use L0 level 'i' for the MC.  This is a minor effect.  Note, the cross section was not computed for new 1 & 2 because of this.

---

Conclusions

• The (SMD u+v)/E cut at 0.006 has almost no effect--might as well drop it.
• The number of clusters per layer cut removes statistics without much benefit--don't use it
• Preshower cut--causes significant impact.  Open question what threshold to use, but something smaller than 40 MeV seems appropriate.
• SMD Seed E--causes significant impact.  It makes the signal shape narrower and reduced the low mass background.  It does cut statistics overall though, so the exact positioning of this cut needs optimization.
• For now, "new 1" seems like the best alternative to "old", but a little more tuning can be done.
• The cross section doesn't change all that much with the various cut options.