Structural Cosmology: A Baseline and Normalization Audit of the v2.2 Demonstration Code

This note audits the baseline and normalization of the demonstration code that accompanies Structural Cosmology v2.2 (10.5281/zenodo.20757059). It examines whether the code reproduces the stated prediction of −5 to −7 % for Δ(fσ₈) near z ≈ 1–1.5 (v1.3 and v2.0; v2.2 states O(5 %) and describes it as consistent with v1.3).

Four pre-registered checks (PR1–PR4) and a reading record found (i) a difference from an approximate GR baseline in the original comparison, (ii) a normalization error of the growth rate in solve_growth, and (iii) that, after the correction, neither of two explicitly defined normalization conditions reproduces −5 to −7 % at the 16 registered points (condition A: maximum |δ| 0.09765 %, class MIXED_SIGN; condition B: maximum |δ| 0.66156 %, class ENHANCEMENT; the quantity is the demo's fσ₈ proxy). The normalization condition used by the original statement could not be confirmed from the materials examined.

Not done: independent re-execution of the numerical results, and any comparison with observations. A non-reproduction by the demo is not treated as a refutation of the theory as a whole. Correspondence of the examined files with the public Zenodo files was compared by SHA-256 and MD5 (9 of 9 files agree). The record includes a materials archive (code, corrected copy, diff, pre-registrations, drivers, outputs, reading record, comparison record, tables and earlier drafts).