Koide Is Not a Face Balance: P8-B's Refusal, with the Corpse Measured

P8-B run to its verdict: is the Koide combination Q = 0.6666605115 (2/3 to a relative defect of -9.2e-6, PDG pole masses) a two-face balance object of this frame? Forty owned forms graded at a tolerance pre-registered at the defect's own scale (6.2e-6): all sever-pair ratios at both members, all two-face splits of the lepton sector under mass and sqrt-mass shares, mass ratios against owned counts, and the Koide angle (0.222270 rad) against the owned small angles. None lands; the closest (T/(1+T) at m = 211, 0.660772) misses by 950x the tolerance. The balance x/(1+x) at x = 2 is shown to be the target's own shadow (x solved from Q is 1.999945, its miss exactly the defect reparametrized) and no owned lepton split supplies its required 2:1 ratio (actual: 73.35, 4.17, 0.26). m_mu/m_e = 206.76828 stands +0.62% from the fixed split 205.5 — refused under the coincidence discipline. Verdict: the relation is real at 1e-5 and is not one of the frame's objects; the two-face address is closed with the corpse measured.