Each work is a pair. The left panel is a photograph as it was made. The right panel is the same photograph carrying one intervention derived from its own measurements. Side by side, the difference often lives at the threshold of vision, and it is stated exactly, down to fractions of a single tonal step. The numbers are the work, and anyone may measure them again. You are offered almost nothing to see, and everything to check. · · · Every drop already inverted the world. Each one now inverts the part of it that is actually behind it. A drop on glass can act as a converging lens, forming a real inverted image of whatever it faces. Measured across 1,904 drops, the ensemble does exactly that in aggregate. The drops run darker at the top while the scene runs brighter above, sky up in the world and sky down inside the water. Per drop, though, it was a fiction. The correlation between what each drop showed and what actually stood behind it was −0.0495, which is essentially zero. Only 56% inverted with the correct sign. Every drop showed the same generic pattern of dark above and bright below, whether it hung over the treeline, the green field, or the ochre bank. They inverted a world, not the world. Each one has been given, under the same flat inversion the ensemble already obeys, the scene it is actually in front of. Measured from the finished file, the per-drop correlation goes from −0.0495 to −0.8792, the share correctly inverting from 56.0% to 91.8%, and, restated as the check's signed top minus bottom, the drops go from −9.16 to −9.51 against a scene value of +10.12, which the intervention leaves untouched. The cost is 55.8 levels at the worst pixel and a mean of 1.36 across the frame. This work's first version was corrected after its numbers were found to have been computed from the prediction rather than from the artifact. The figures above are from the file, and that correction is part of the work's record. Presentation. The mount is neutral 150, the drops' own mean luminance of 149.7. The full derivation gives the drops' mean color, a warm khaki, but a chromatic surround visibly corrupted the photograph's own warmth, so the water sets the value and is denied the tint. There is no rule. To check. 133,373 pixels move, none by more than 56 of 255. Download this work's original file, 06_real_image.png, not a screenshot or a marketplace preview, attach it to an AI that can run code, and paste this: Run this exact line on the attached file, changing only the file path if needed. Show me its exact output, and say whether that output is exactly: peak 56 changed 133373 python3 -c "import numpy, PIL.Image as I; c=numpy.asarray(I.open('06_real_image.png').convert('RGB')).astype(int); a=c[797:3455,797:3455][::3,::3]; b=c[797:3455,3854:6512][::3,::3]; d=abs(b-a); print('peak', d.max(), 'changed', (d.max(2)>0).sum())"