The Surface From Broken Bonds — the obvious derivation attempted, its miss measured, its author corrected

Paper 80 named the surface coefficient (17.8 MeV) as the only one of the four with no derivation anywhere. The direction lists asked for the obvious attempt: compute the surface energy from packing and coordination. Done, exactly — FCC clusters built ball by ball, broken bonds counted, the bond energy fixed by the volume term (, nothing fitted) — and the answer is 21.4 MeV against the fitted 17.8: a zero-parameter classical estimate that lands 20% high. Two honest records ride with the number. First, the author's expectation was closed by the run: the pre-run reasoning said a broken-bond count must land below the volume coefficient; the computed count says otherwise, and the runner's own docstring prints that result as it came out rather than smoothing it away. Second, the direction of the miss teaches something: the classical count over-charges the surface — reality is about 20% cheaper than missing-neighbour arithmetic — and the named candidate for the discount is quantum wall-softening (the wavefunction spills past the last row of balls, so a surface ball is less starved than the bond count says). That leans the same way as Paper 79's finding that the surface law overcharges small drops, and Paper 86 finds the same diffuseness on the radius scoreboard. This record holds one paper on its own and is refreshed every couple of months. The complete deposit, 10.5281/zenodo.21788037, holds every paper with its code and data and is updated first; for the newest version of this paper, go there.