After looking at BeamLine vs. Luminosity in Run 12 UU193, I decided to have a look at Run 11 AuAu200. So I spun through 750 MuDsts (2500 events from each), from fills 15999, 16000, 16002, 16004, 16005, 16006, runs 12153032 through 12155040, using only the protected minbias triggers "vpd-zdc-mb-protected (offline trigger ID 350033 for runs from the first two fills, and ID 350043 for runs from the last 4), with a cut on rank > -3 (just as with the UU data).

There is some mild structure within the fills visible:

Here I show the luminosity (ZDC coincidence rate) dependencies of x0 and y0, with fill 16002 highlighted in red, as it seems to behave differently from the other fills:

There does appear to be some luminosity dependence, but it isn't linear. Here is the same information in profile (slightly smaller ZDC range to stay where I had good stats):

So the observed variation with luminosity is of order less than 100 microns over a wide ZDC coincidence range, which is much smaller than the currently observed variations of order more than 1 mm (taking into account both x and y) in Run 12 UU data.

Just an aside (not absolutely certain it is relevant at this point), but the TPC luminosity-dependent SpaceCharge & GridLeak calibration for this data was a linear function of the BBC sum. Fill 16002 may have had something amiss with using the BBC, or with the beam in general. Here is the BBC sum / coincidence ratio vs. time (arbitrary zero) for these 6 fills, demonstrating notably different behavior for fill 16002:

-Gene