UDCT v7.12 Post-registration Results: Anisotropy and Screening-Length Boundaries of Signed Jeans Solutions in Hydrus I

Abstract

This record reports the completed UDCT v7.12 preregistered numerical experiment on anisotropy and screening-length boundaries of signed radial Jeans solutions in Hydrus I. The frozen protocol compares five screened GQUMOND model variants and three QUMOND controls, using constant velocity anisotropy β from 0 to 0.5, screening lengths from 1 to 3 pc, and predefined grid and inner-boundary comparisons. The fixed matrix contains 363 primary decisions: 274 negative (N), 48 positive sampled candidates (P), and 41 indeterminate (I). The screened subset yields N/P/I = 274/23/33, while the controls yield 0/25/8. All five screened models are classified N at 2.5 and 3 pc throughout the tested anisotropy range. Some shorter-length, higher-anisotropy settings retain model-dependent P candidates. Of 575 adjacent intervals examined, 24 contain strict raw sign changes. All 24 remain unresolved under the registered numerical-margin stopping rules; no exact critical anisotropy or screening length is established. Including adaptive evaluations, 246 fields, 2,286 pressure records and 381 case decisions were completed. All 246 isotropic reproduction gates passed. The advance is a denser, reproducible map of conditional Jeans compatibility, with explicit model differences and unresolved boundaries. Negative radial pressure indicates incompatibility with the specified stationary Jeans closure. Positive second moments alone do not establish a nonnegative distribution function, equilibrium, dynamical stability or observational agreement. These results neither establish a universal Hessian threshold nor generally exclude MOND. The report discusses future numerical-resolution, distribution-function, stability and observational tests. These proposals were not executed in this experiment and require separate prospective specifications for confirmatory follow-up. The upload includes the linked report PDF, README, a Python launcher, and a lightweight ZIP containing the unchanged scientific source files, pinned requirements, case and bracket tables, checksums, and execution/timestamp evidence. The full compressed raw output and incremental journal are excluded from this lightweight package and retained separately. The report distinguishes the original observed-public-status timestamp bound from subsequently verified Zenodo record timestamps. Results repository:https://github.com/wspaik5/UDCT_v7.12_Postregistration_Results DOI reserved for this results record:https://doi.org/10.5281/zenodo.23296444 Public preregistration:https://doi.org/10.5281/zenodo.23280972 Previous v7.11 results:https://doi.org/10.5281/zenodo.23263099