After fixing the L2 thresholds in the trees, I still noticed more events in the lowest p_T bins. As it turns out, the main TSIU cluster find class was changed to require N_strips ≥ 5 rather than N_strips > 5. It appears the original intent was to include 5-strip clusters, so we may prefer the current trees to the old. However, before we move toward a preliminary result we should asses how the situation is changed for including the 5-strip clusters. To this end, I have calculated the asymmetries for the two samples. This required reproducing Part-2 and Part-3 of the Monte Carlo trees. I have calculated the systematics as we discussed at the ANL face-to-face meeting. The signal-fit-parameter uncertainties are included solely in the statistical uncertainties; and the alternative fitting procedure and the non-collision systematic have been removed from the calculation.

Figure 1

Nstrips > 5	Nstrips ≥ 5

Qualitatively, the two samples are similar. Perhaps, the biggest change is that the systematics are larger in a few bins (e.g. x_F > 0 bin-2) for the new trees.

Figure 2

Nstrips > 5	Nstrips ≥ 5

Again, the p0's are qualitatively similar between the two samples.

Table 1: Old Trees, Old Run List

Table 2: New Trees, Old Run List

I have fit the p₀ distirbutions with a constant. The fits are all well within 1σ of 0.

Table 3: Old Trees, Old Run List (x_F > 0)

Table 4: Old Trees, Old Run List (x_F < 0)

Table 5: New Trees, Old Run List (x_F > 0)

Table 6: New Trees, Old Run List (x_F < 0)