PulseExploreJournal ClubDebatesTrendingResearchersJournals
Instagram
HomeExploreJournal ClubTrending
Synapse
⌘+K
Synapse
May 17, 20260 citationsOpen Access

Robustness Tests of a Foundation-Only Low-Redshift Directional Residual in DES-SN5YR

View Full Paper
MHMalin Hess

Key Points

  • This analysis aims to assess the stability of a low-redshift directional residual within the DES-SN5YR dataset.
  • Analyzed 92 objects with zHD<0.060 from IDSURVEY=150.
  • Conducted a series of internal robustness tests, including bootstrap and leave-k-out resampling.
  • Performed exhaustive leave-two-out tests covering 4,186 object-pair removals.
  • No object pair removal reduced Δχ² below 10 or shifted the fitted direction by more than 15°.
  • Minimum leave-two-out Δχ² was 10.947866; median was 14.740262; maximum direction shift was 10.795520°.
  • The residual remains robust under the applied diagnostics, albeit with caveats regarding velocity-frame sensitivity.

Abstract

This archive presents a robustness note and supporting diagnostic package for a DES-SN5YR Foundation-only low-redshift directional residual analysis. The primary subset is IDSURVEY=150 with zHD<0.060, containing 92 objects. The baseline covariance-aware directional residual fit gives Δχ²=14.945456, amplitude=0.101681 mag, and best-fit direction RA=20.366097°, Dec=57.880151°. The analysis tests whether the Foundation-only low-z feature is stable under internal robustness diagnostics, including random-sky nulls, residual sign-flip nulls, bootstrap and leave-k-out resampling, redshift splits, local-flow/VPEC cuts, velocity-frame sensitivity tests, leave-one-object influence, and exhaustive leave-two-out object-pair removal. The exhaustive leave-two-out test covers all 4,186 possible object-pair removals. No pair removal reduces Δχ² below 10 or shifts the fitted direction by more than 15°. The minimum leave-two-out Δχ² is 10.947866, the median is 14.740262, and the maximum direction shift is 10.795520°. AT2017zd is identified as the leading leverage-heavy object, but it is not result-defining. This archive is intended as an internal robustness and reproducibility package. It does not claim cosmological anisotropy, a preferred cosmic axis, URSF confirmation, independent object-level replication relative to Pantheon+, or elimination of local-flow/velocity-frame systematics. The strongest defensible interpretation is that the Foundation-only low-z residual is internally robust under the tested diagnostics, while its physical interpretation remains caveated by velocity-frame sensitivity, local-flow treatment, and complete low-z Foundation object overlap with Pantheon+. SHA256: 3f61f556960ecd4a3c9c5a2eee0ff560c88360b1ba93fee50162d5fb72a79e9b Prepared by Malin Hess with assistance from OpenAI’s ChatGPT for code-structuring, documentation, formatting, archive organization, visual generation, and wording review. Scientific interpretation and final responsibility remain with the author. All code is under MIT licensing.

Ask AI
Helpful
Bookmark
Share
View Full Paper

Cite This Study

Malin Hess (2026) studied this question.

synapsesocial.com/papers/6a095c6d7880e6d24efe285ahttps://doi.org/10.5281/zenodo.20206067
Ask AI
Helpful
Bookmark
Share
View Full Paper