I'm using Jan's Algo ver 4.1 with a few modifications to look into the idea Scott suggested of sorting events by placing conditions on the away side to get an estimate of the background.  Here is a list of the cuts used:

CUTS ver 4.1 algo params: trigID: bht3=230531 L2W=230601 BHT_Dsm>28 isMC=0
TPC: nPileupVert>3, vertex |Z|<100.0cm, globEleTrack: nFit>15, hitFrac>0.51 Rin<90.0cm, Rout>170.0cm, PT>10.0GeV/c
BTOW ADC: kSigPed=3 maxAdc>200 clustET>15.0 GeV ET2x2/ET4x4>0.9 ET2x2/nearJetET>0.8
dist(track-clust)<7.0cm,
W selection highET>30.0 awayTotET<8.0GeV

I relaxed a couple of the cuts from Jan's analysis:

ET2x2/ET4x4>0.9 ET2x2/nearJetET>0.8

We probably get a few more W's and a lot more background, but you can still see a bit of a peak in the lower right of Fig 1 (where no away side requirements are in place)

Figure 1:  2x2 Cluster ET distributions for different away side conditions (conditions shown in title)

I think it is encouraging that when we see something "jet-like" (more than 2 tracks and some reasonable energy) on the away side the Jacobian peak essentially disappears.

Figure 2:  Since the upper left and lower left have about the same y-scale I plotted them on top of each other here, no normalization is applied.