Everest II Outcome — The Seventeen Have Spoken: Consolation Tier, the Reversed Head-to-Head, and the Post-Mortem (The River and the Whirlpool)

ROUND-5 RESTATEMENT (metadata revision, August 5, 2026, per referee finding R5-1b): every reference to "10%-floor errors" in the enclosed documents denotes, verbatim, σᵢ = √(errVᵢ² + (0.10·V_obs,i)²) — ten percent of observed velocity added in quadrature; a literal max() floor was never used. The adequacy PASS (χ²/N = 1.136 vs 1.5) is specific to that instrument; under a literal floor the candidate would have failed the bar and the killed RAR passed it. The A4 head-to-head that decided the outcome is a ratio and holds on either reading: the candidate loses to the RAR, and the CONSOLATION-TIER verdict stands. See PROGRAM CORRIGENDUM (August 5, 2026), which also carries the referee's Erratum 2 (arXiv:2008.04065 is Zhou, Del Popolo & Chang 2020; Marra, Rodrigues & de Almeida 2020 is arXiv:2002.03946). VERSION NOTE (v2): Version 1 of this record was published with an incorrect file set — the author uploaded the adjacent Everest II FREEZE folder (duplicating DOI 10.5281/zenodo.21808357) rather than the outcome files. Per the program's no-overwrite and full-disclosure rules, v1 stands as published and this version carries the correct outcome file set. The error was the upload step's, not the record's; nothing scientific changes. The complete record of the single certified blind run of the derived zero-knob candidate mechanism on seventeen untouched SPARC galaxies, executed August 5, 2026, hours after the registration was frozen (DOI 10.5281/zenodo.21808357), and published in full regardless of outcome per the certified §8. Order followed exactly: forward discrimination first — all certified bars met before any observed velocity was read (decisive candidate-vs-RAR separation 273.3 against the required 100) — then one unblinding, then adjudication strictly per the certified §5. Verdict: CONSOLATION TIER. The candidate passes the absolute adequacy bar (joint χ²/N = 1.136 on 344 points, bar 1.5) with both kill tripwires clean — and fails the registration on amendment A4: it does not beat the killed RAR-form mechanism head-to-head under matched instruments (1.136 vs 0.980), and the free-Burkert comparator cap fires (ΔBIC +976, candidate charged k = 1). The spent-set head-to-head that led the freeze deposit (1.78 vs 2.07) REVERSED on blind data and is retired from all future citation. The pre-registered spent-set inner-zone bias did not persist (z_inner: +4.89 spent → −2.71 blind); a new outer-zone systematic is reported (z = +5.36 beyond 3r₀). The certified gas-rich blind pool is exhausted at 29 galaxies. Post-mortem annex (hindsight-grade, zero evidential weight, labeled): the strength-dial diagnostic (a ×2 / ÷2 effective-coupling conjecture refuted; the data demand ordinary G for the dressing, supporting the component reading of record); the halo-archaeology memo (the 29 free-fitted halos reproduce the Donato column's central value, 129 vs 137 M☉/pc², but refute its universality at ~4σ on our own sample; no present-day baryon observable determines these halos to better than ~0.22 dex; the best possible zero-knob law of the family scores 2.76 against a free floor of 0.15 — the family, not the member, is the limit); and RD-9, the Two-State Budget Wager, the surviving cosmological form of the coupling conjecture, parked for the cosmology front. The program publishes its losses at the same grade as its wins. Filed by the author with AI research assistance (Claude, Anthropic); all claims and errors are the author's. Prior art credited throughout the enclosed documents. The 2020 original remains untouched.