I think you'd better find some different examples Tony. In the case of the shot you included, the nearest-pick looks a little better than the empeg optimized one.

And like I said, the important point wasn't which one looked better, but that the two used different indices to derive their cutoff points.

Personally, I preferred the empeg-optimized one for various reasons, including things like the details in Cameron's jersey were more visible, the expression on Ferris' face is more clear, etc. But those are simply subjective judgements and have nothing to do with the discussion.

And if you're dithering, it becomes even more critical, by the way, because the shade of each dither dot becomes more important.