# n226 — The survival ledger: every predecessor claim, re-audited sequentially against P3″

**Date:** 2026-07-11 · **Method:** two passes. Pass 1: a 31-agent parallel workforce (86 claims audited, 24 flagged, each flag adversarially verified, several by fresh computation — full record in the workflow journal). Pass 2, at the human's direction: **a sequential re-audit by the builder**, slice by slice, with the workforce as second opinion — the two decisive computations re-run by hand (the multi-pole knee shape; the residual-force glass check confirmed from the verifier's script). Where the passes disagreed, the sequential result governs and is noted.

## Scoreboard (sequential-final)

| Verdict | Count | Note |
|---|---|---|
| SURVIVES | 31 | topological / statistical / graph-generic / substrate-free reasons only |
| RESCOPE | 36 | true with a changed statement — all applied or queued below |
| AT-RISK | 9 | named recheck each; sharpest: entry 30's charge-blindness under two ledgers |
| BROKEN | 3 | corrected on-site this pass |
| ALREADY-DEAD | 7 | pre-transition retractions, confirmed still marked |

## The three BROKEN, corrected

1. **The DM–knee cross-lock (entry 28 / n187 channel 1): CUT.** Its left side was the demoted capacity band (§§40/43); no amorphous winding-capacity model exists. The quench clock survives conditionally (σ ∈ [0.50, 0.62], δτ/τ ≈ 1.8·δm/m).
2. **n211's "knee within 0.1% of Lorentzian": corrected by my own re-run.** The 0.1% was the N = 4 ring; realistic substrates give **~6% on crystal AND glass alike** (5.8% / 6.1% — a graph-size fact, not a disorder fact; the workflow's glass-only 5–9% is thereby sharpened). Inside the two-pole discriminability band; registered as a native prediction. Pole identity stands graph-generically.
3. **n197's projector theorem: museum piece.** Its setting (exact crystal momenta, derived cubic weights) left with the substrate; its program died with n212/n215. Status page now past-tenses it. The owed glass computation (watched/dark decomposition vs seed) is registered.

Plus two residuals the sequential pass caught that the site still asserted: the §49-retracted "clock-invisibility IS gauge invariance" surviving in the dichotomy entry (rescoped: the dichotomy is now the structural reason the two ledgers exist — the theorem was the split introducing itself); and the "one-stroke web" (parameters) suspended — its last link routed through the demoted single-quantum band and P3″ splits the sectors it chained.

## The pattern (the audit's real yield)

What SURVIVES divides cleanly by reason-class: **topology** (photon kernel dim E−N+1; windings as homotopy classes; |q| ≤ m as a triangle inequality; matter-cannot-leak's surgery argument), **statistics** (polarization equality; achirality as ensemble statements), **graph-generic algebra** (dark-state derivation; pole identity; the IBP/η_N identity; the force law's constraint-owned pole), and **substrate-free mediator results** (ζ = 1/4; odd/even, prediction 33; the passivity and gradient theorems). What BROKE or rescoped heaviest: anything that lived in **crystal momentum space** (projectors, direction-set weights, staggered-point identities) or chained sectors through the **single cost scale** (the anchor's cross-sector derivation, Q's four jobs, the consistency web, the de Broglie four-ledger phase via n166's crystal rows). One sentence: *the theory's truths were always the topological, statistical, and algebraic ones; its crystal-exact statements were scaffolding, and the scaffolding is what the transition stripped.*

## AT-RISK queue (each with its named recheck)

30's charge-blindness (re-run n171's reductio with sector-split λ_e — sharpest, gates entry 36c); the knee anchor's two-ledger re-derivation (entry 2); n199's momentum-billing on the glass (needs glass flux rows); entry 20's per-realization antimatter exactness (cross-sublattice coupling check; BASE becomes a live constraint if ensemble-only); the R = 1 numeric on glass; collisionless amplitude with amorphous spacing; preferred-frame dynamical margin under P3″; the KZ → η link (does the quench produce the disorder it needs?); Q's job-by-job sector re-audit.

**Class: ledger [record of record]; corrections [applied, rule 5]; the workforce journal is the parallel record, this note is the sequential authority.**