BB-CBE-3b: Almost-Everywhere and Centered-Spike Architecture

This project evaluates the first footprint-aware, centered-spike successor to the unpolarized Gowdy test. It preserves three production resolutions, fitted diagnostics, figures, and an independent verifier.

Result

3b-D — architectural nonclosure; no physical equivalence verdict.

• The original 99.9%-mass footprint saturated almost the entire circle and therefore failed as a feature localizer.
• A P-only window velocity retained a convergent drift, while the preregistered hyperbolic-velocity variant was substantially more stable.
• Forward prediction remained intermediate rather than conclusively passing or failing.
• The Lim fit was not a valid membership test because its training mask spanned the whole inhomogeneous field.
• Independent verification passes all 32 checks.

Contents

• run_cbe3b.py — production test.
• verify_cbe3b.py — independent recomputation of gates and outcome.
• CBE3B_RESULTS.json, CBE3B_VERIFICATION.json — result records.
• cbe3b_slices_N*.npz — preserved resolution data.
• fig_*.png — diagnostic figures.

Reproduction

Clone bb-cbe3 and bb-cbe3t beside this repository. Requires Python 3, NumPy, and SciPy.

python3 run_cbe3b.py
python3 verify_cbe3b.py

Scope

The outcome is a test-architecture failure, not evidence for or against physical collapse–bang equivalence.