Aside from this, there is a trivial shift in the y-coordinate for all hits arising from our model of the data.
Below is a plot of the hit positions on padrow 128 for one event at 2 A GeV. The hit positions from the first pass are the black circles, the positions from the second are the red stars. The green circles are just the first-pass positions shifted by -2.2 cm in y. Clearly, for this picture, the initial guess at position leads directly to the final position, modulo an offset in y.
Some hits (see above) do get shifted (on a scale hard to see in the large picture) between
first and second pass. The difference between x-coordinate for refit hits is shown on the
next plot. Keep in mind that a pad's width is 0.8 cm, so the amound of "wiggle" imposed
by the fit is only about 5-10% of the potential "digitization scale."
For the y-position, on the other hand, the amount of wiggle imposed by the refit is on the order of one full bucket's width (~5 mm). This is clear from the second panel of the following plot.
[In both the x and y plots, there are 70/1151 hits which do not get shifted at all. These
are mostly artifact hits at the bottom of the TPC.]