Iteration 06 of the GCM + O-information orthogonality-selection loop (S0–S5).
WATCH for GCM, WATCH for O-information. Both are correct (S1), calibrated (S2), reproducible & leakage-free (S3), and stable (S5) — but value-add over the production baseline {Pearson, Spearman, h_norm, MI} is NOT demonstrated (S4), because the only non-circular ground truth is itself Pearson-visible. Confirmed by a 3-lens adversarial review (0 material flaws). Flip to PROMOTE requires a non-empty FDR-survived S4 via a nonlinear/ratio construction identity or a Pearson-independent forward-return answer key. Deployment is Terry-gated. Full verdict page: instruments/gcm-oinfo-selection-verdict.html.