← Phase 1 / 22 · v0.5 Exact Census Report
Within the three-stage strategy defined in document 21, Gate A (delta-only base verifier) and Gate B (twist JSON parser diff) are truly executed here, yielding precise numbers—the content of this page is located in the suite's results/ directory rather than docs/, making it one of the few documents in this series without a docs/NN number, but it belongs to actual research output rather than intermediate scaffolding, so it is still included separately. The document first reveals a provenance detail: the OLD twist JSON was actually frozen before the OLD base file (originating from an earlier generator commit), resulting in 1355 OLD base curves having no corresponding twist entries; the document explicitly points out that this is an "archived output artifact," not a fresh end-to-end execution on the OLD source tree. Core precise results: base curves are old=40,749, new=36,687, removed=4,062, added=0; Algorithm 1 failure classification ISOGENY_ONLY=1,353, A3_ONLY=2,707, BOTH=2, UNEXPLAINED=0 (individual: 3-isogeny 1,233, 5-isogeny 115, 7-isogeny 7); the precise change for Algorithm 2 in the stable-domain is a pure reduction of removed=21,306, added=0, with curve-level classification UNCHANGED=31,250, SHRINK_ONLY=5,437, EXPAND_ONLY=0, MIXED=0; global twist accounting identity 46,091=46,091 exact PASS. All 26 items on the completion gate checklist are true.
Relationship with other documents, try to use the words from its own document, not my interpretation.
"This is an exact archived-output census of a theorem-producing computation. It is not a fresh end-to-end rerun of every OLD curve, and it is not a proof of the Birch–Swinnerton-Dyer conjecture for all elliptic curves." — Excerpt from the conclusion of this document.
Loading...