Reproducible Low-RMS E8 Orientations Under a Fixed-Catalogue Rotation Null: A Pre-Registered Test and Surrogate-Catalogue Check (Version 4.0.0)

Version 4.0.0 of the E8 lattice alignment investigation reports a pre-registered predictive test of the version 3.0 low-RMS rule on fresh projection seeds, together with a supplementary surrogate-catalogue check that materially bounds its interpretation. A 1,500-seed tournament was run over seeds 300–1799, a range disjoint from all earlier versions. The elite threshold, a refined RMS below 0.047°, was set in version 3.0 prior to the tournament. The refined shortlist contained 16 observed threshold-crossing seeds. A pre-registered random subset of 12 was evaluated using the version 3.0 search-aware rotation-null protocol, with 1,000 best-of-search null trials per seed. Ten of the twelve tested seeds had empirical (p<0.05) and were rejected by the Benjamini–Hochberg step-up procedure applied within the tested 12-seed family. None of six pre-registered above-threshold control seeds was significant. This constitutes a pre-registered predictive validation of the version 3.0 low-RMS selection rule under the fixed-catalogue rotation null. It does not constitute FDR control over the upstream 1,500-seed tournament screen. A supplementary pre-registered longitude-permutation pilot tested whether the low-RMS tail was specific to the catalogue’s real latitude–longitude pairing. Four surrogate catalogues preserved the observed latitude and longitude marginals while randomly permuting their pairing. Under the identical 1,500-seed pipeline, the surrogates produced 14, 21, 22 and 24 seeds below 0.047°, compared with 16 for the real catalogue. In descriptive post-hoc tail-depth summaries, all four surrogates achieved a lower minimum refined RMS and a lower mean among their ten best seeds. With only four surrogates, no catalogue-level permutation (p)-value is reported; the result indicates that the real coordinate pairing shows no observed advantage in this pilot. The combined result is interpreted as a reproducible interaction between the E8 network geometry, the preserved catalogue marginals, the scoring statistic and the many-seed search—not, by itself, as evidence of an Earth-specific E8 alignment. Future work will investigate which preserved structural feature produces the surrogate tails and whether frozen v4.0 orientations predict independent data once that effect is controlled for. The deposit contains the paper and LaTeX source, signed and OpenTimestamped pre-registration documents, the realised seed-selection record, tournament outputs, eighteen rotation-null result sets and logs, scoring and audit scripts, four longitude-permuted catalogues with their tournament outputs, the five-catalogue summary, integrity manifests and citation metadata. An interactive globe viewer is available at: https://salahealer9.github.io/e8-globe-viewer Archived viewer version DOI: 10.5281/zenodo.21429524 This version supersedes version 3.0.0: 10.5281/zenodo.20759785