I updated the correlation macro to loop over z as well as energy and angle.
Fraction of sky covered by AGN (increases with z cutoff):
The first set of maximum correlations was found for z=0, with the next best at z=.002. I think that the z = 0 correlations are not to be taken very seriously because they basically represent one data point which happens to line up with an AGN. Because the fraction of sky being considered in those cases is so small, the expected number of correlated data points is ~0.02. Thus an accidental line up of 1/31 data points looks very significant. You can see a very big jump at E=41 EeV, when the data point falls below the threshold.
Here are the plots of the scans at z=.002. The best set of parameters has a match of 5/22 data points.
The next best correlation was at z=0.018, which was the redshift found by the initial Auger scan. This is plotted on April 16. It matched 16/27 data points.
Root file: correlation3.root.