Quantized Horizon Response (HRF): Composite-Alternative Separation and Fixed-Conditioning GW250114 Replay - v5.3

Abstract

Methods and reproducibility release. Research package v72; Zenodo publication version v5.3. This release makes available the corrected GW250114 HRF composite-alternative separation methods, numerical design calculations, and the archived fixed-conditioning -7M raw-grid replay, together with a September 2026 publication audit and reader guide. The original v72 archives, sealed on 17 August 2026, are preserved byte-for-byte. In More Basic Terms After two black holes merge, the resulting black hole settles down through gravitational waves, somewhat like a struck bell ringing down. This project investigates whether a proposed horizon-response model could eventually be tested with those signals. The archived calculation closely reproduces a reference grid when specified assumptions are held fixed, but changing the released parameter row shifts the best-fitting damping rate. Reproducing the calculation therefore does not, by itself, establish a new physical effect. The statistical work explains why a finite measurement cannot uniformly distinguish one exact value from every arbitrarily close alternative under the stated continuity condition. A meaningful test needs a declared resolution, an interval procedure, or a specified Bayesian comparison. The progress is a more reproducible test and clearer limits on what can be concluded. It does not establish a unified theory of everything or a quantum structure at a black-hole horizon. This is an independent exploratory research package, not an official LIGO/Virgo/KAGRA product. Technical scope Continuous-alternative closure: under the stated total-variation continuity assumption, the exact point target has zero uniform separation from the punctured continuous alternative; the stated randomized minimax maximum error is 1/2. Least-favorable Gaussian design: one named competitor requires distance at least 3.7810604375 for target-side score exceedance of ln(10) with probability at least 0.90. The corrected symmetric two-endpoint rule requires at least 4.3486867653 per endpoint for a 90% joint-success lower bound. Effective-width screens at the historical gap are 0.115190902% and 0.100155239%, respectively. These are design calculations, not event measurements or automatic composite Bayes factors. Archived fixed-conditioning replay: the 85 x 40 -7M grid has a released-row network maximum at 200 Hz and 450 inverse seconds, with centered cache-to-raw residual below 7.8e-5 log-likelihood units. Selecting another released posterior row moves the damping maximum to 420 inverse seconds. Active nuisance parameters change together. The September audit verifies supplied artifacts and distinguishes fresh local checks from preserved raw-replay receipts. Official raw inputs are not bundled, and this publication preparation does not claim a new raw-strain replay. A documented prose erratum corrects the recovery audit's signed residual label: the negative mean offset is cache minus raw, not raw minus cache. Centered residual bounds and gate dispositions are unaffected. Limitations and next work The full reference inference and its marginalizations are not reproduced. Detector-only remnant/QNM inference, global nuisance-image separation on a real joint likelihood, uncertainty propagation, and full-scan null calibration remain incomplete. Both detectors' event data have already been examined, so the accompanying proposed detector-disjoint reanalysis is a draft protocol, not a retrospective blind holdout or an executed study. No HRF detection, exact ln(4) or g=4 identification, area quantization, black-hole microstate inference, or quantum-horizon discovery is claimed. The deposit preserves methodological progress without changing these physical interpretation limits. Files and provenance Begin with the reader guide and dated publication audit. The focused v72 ZIP contains methods, code, numerical surfaces, receipts, and manifests; the master ZIP preserves cumulative research history. SHA-256 sidecars and a publication file manifest support integrity checks. Historical reports retain their original July/August dates. September documentation and local checks were prepared with AI assistance in Codex and do not constitute independent peer review.