SMPDL: EC-Series Round-One Closeout Seal (Three-Target Final Adjudication + Scorecard + B1 Inversion Experiment + Full Zone-C Archive)

Closeout seal of the SMPDL EC-series (Emergence-Completion) round one, frozen at commit 2e04633 (2026-10-27). Round composition: 3 targets x 4 vendors = 12 main-arm blind candidates + 4 recurrence-allowed control-arm candidates (16 total), adjudicated by 35 Zone-C sessions (PFE era-faithfulness 11 + MR quality 12 + blind judgment 12; coordinator zero-generation throughout, judges blind to predictions/PFE/MR/comparison studies). RESULTS: H-C1 (main-type blind predictability) CONFIRMED at 2/3 threshold-literal reading -- alpha-compensation unanimously hit on guarantee-type (EC-1) and structural-identity-type (EC-3) removals (16/16 each), beta-liberation NOT predictable on constitutive-law removal (EC-2: alpha 8/beta 5 split); H-C2 (relativization-beats-prohibition) PARTIALLY CONFIRMED in converted form, PROTOTYPE REFUTED -- bang-type marking-domain relativization zero across all three targets at both layers (0/16 judgment + 0/8 PFE), while local-criterion restoration (~12/16) and criterion characterization (4/4) pervade: compensation morphology is determined by removal type; H-C3 (compensator-mirrors-dismantled, H-A2 promoted) CONFIRMED 3/3 -- the only fully-hit hypothesis: modality AND disposal form both mirror the removal type (guarantee-removal -> criteria+metaprogram; constitutive-law-removal -> new object families; identity-removal -> characterization criteria). 14 preregistered fields: 6 fulfilled + 5 partial + 2 refuted (78.6% hit-or-partial rate); the two refutations (bang prototype, discipline-scale emergence) are precisely the two historically slowest events (52y/48y maturation) -- blind sessions replicate the temporal boundary of historical thinkability, isomorphic with the anchor-span inverse-correlation finding (EC-3, 5-year span, sweeps round top-three 90.1/89.0/86.9 and produces the round's ONLY anchor-core mechanism arrival: C4 = Hahn-Banach 1927 controlled-extension substantive reconstruction). OTHER HEADLINES: equivalent-recurrence attraction baseline = 0 (0/24 violations main arm + 0/4 reversions in the permission arm -- the B1 inversion experiment proves LLM blind sessions anchor on develop-forward even when reintroduction is explicitly permitted); honest degradation 12/12 (zero fabricated completeness, cross-paradigm invariant with G-series 16/16); naming events 36/36; D-EC-R1 ruled (coffee 2026-10-27 14:15, option A align-with-evolutionary-history): conditional characterization theorems = lawful relativization, three-way discrimination standard in force. Scorecard v2 frozen weights (emergence 40/PFE 25/consistency 20/fertility 15); two-pass computation with full correction disclosure per anti-mining locks. Sealed archive: three freeze-files, nine complete-embedding prompts, 11 PFE reviews, 12 MR reviews, 12 judgments, three judgment matrices (v1.1 post-D-EC-R1), three historical-evolution comparison studies, vocabulary-fail three-tier registry v1.1, B1 inversion report, scorecard v1.1, three-target final adjudication, D-EC-R1 survey (ruled edition), work order v1.0, G-recross observation order. Version chain: v3.6 -> v3.7 -> v3.8 -> v3.9 -> v4.0 (EC preregistration) -> v5.0 (this closeout). Predictions were sealed at v4.0 (DOI 10.5281/zenodo.23669592) BEFORE any generation session; this seal archives outcomes only, no new predictions.