Bob suggested that the curious increase in 3-hit stubs in odd half-cell stacks relative to even ones might be geometric. Look at these plots: the left hand column is even half cells, the right odd half cells. The quantity plotted is the number of times a half-cell had 3-hit stubs. Top is west, bottom is east. As you can see, there are about 4 times as many odd 3-hit half cell stubs as even. I threw together a simple model to see if this made sense.
Essentially I just vary the tangent of a line and its initial position at the level of the 0'th cell, and count the number of hits. There are two cases: even (no offsets between cells due to interleaving gaps) and odd (which have offssets). The black plots are even, the red are odd, and the horizontal axis is the tangent of the angle. In the first plots I use perfect hit efficiencies, and the second set I use Bob's most recent numbers for inner and outer wire efficiencies (.993 and .991). I average the results for positive and negative tangent.
As you can see, geometric effects seem to cause a surplus of 3-hit stubs in odd half-cells. Unfortunately, the model doesn't get the ratio of 3-hit to 4-hit stubs right.
Modified 6-July-2004 at 11:38:14
http://hep.physics.wisc.edu/~jnb/imu/6Jul2004