--- a/research/IMPORT-MAP.md +++ b/research/IMPORT-MAP.md @@ -174,6 +174,7 @@ | 21 | the eigenstructure of the gap transfer matrix, in the cycles-of-gaps convention | `M_J = R · Λ · L` with `LR = I`, upper-triangular entries of `R` and `L` binomial and independent of the prime, and eigenvalues `a_{kj} = ∏_{q=17}^{p_k} (q − j − 1)/(q − 2)` with `a_{kj} > a_{k,j+1}` and `a_{kj} → 0`; Table 1 prints eight values at `p_k = 999,999,999,989` from `p₀ = 13`, `a_{k2} = 0.10206751799779` first among them, and the authors state that convergence of the gap ratios is governed by `a_{k2}` and is slow. Holt and Rudd, *Eratosthenes sieve and the gaps between primes*, arXiv:1408.6002 §5.1, Table 1 and the Figure 3 caption **[SOURCED, verbatim at the arXiv full text, sha256 in the record's §6]** | the fold's histogram operator, `a3-09-histogram-operator.md`, whose own ledger already records it as a rediscovery of Holt-Rudd §5 | none reached; the row is an anchor and a wall address and was never priced as a target | **EXACT-IDENTITY at the operator**, and wrong precision and wrong functional at the conclusion | **CLEAN.** An eigenvalue of a finite matrix with an explicit product formula carries no hypothesis of postulate strength, and it carries no conclusion about the anchor either | PUBLISHED-ANCHOR, which is the strongest available here, plus a WALL-ADDRESS; and a `SEARCH-CONVENTIONS.md` row owed, "cycles of gaps" and "eigenstructure of `M_J`". No THEOREM, no DERIVED-CONSTANT | 1 h | **PRICED 2026-08-29, resolved at recon grade; the novelty-check producer has NOT been written.** Three deciding facts, none of which a sharper analysis narrows. The spectral gap is `1 − a₂ = 1 − 2.82/ln z + o(1/ln z)`, a polylogarithmic rate against killer 2's threshold `ε < 1/W = e^{−(1+o(1))x}` (`attack-wrongdirection-audit.md` §1 Axis C). The functional `a₂` governs is the population ratio `w_{g,1}(∞) = N_g/N₂`, a first-moment count of gaps of each bounded size, and `G₂` is a maximum. And `M_J` is restricted to spans `\|s\| < 2p` throughout, so `G₂` is outside its state space, while `J → ∞` is not a limit of this eigenstructure because `(q − j − 1)/(q − 2)` turns negative for `j ≥ q − 1`. **The anchor is graded below the recon's first draft, and the correction is the red team's:** `history/staging/lit-pdf-holt-rudd.md` line 259 already carries, verbatim and page-numbered, the per-prime eigenvalue `a_j = (p − j − 1)/(p − 2)` for the bidiagonal `M_J`, so what this row adds to the corpus is the product-over-primes closed form, the printed numerical values and the rate, beside the binomial eigenvectors `PRIOR-ART.md` had already attributed (`redteam-0829-measure-c.md` §1h row 1). The rate constant is **[SCRATCHPAD-GRADE]**: `a₂ · ln z` reads 2.816534, 2.819348, 2.820118, 2.820203 at `z = 10⁴` to `10⁷` from a scratchpad producer, reproduced digit for digit on an independent sieve by the red team (`redteam-0829-measure-c.md` §1h row 5), with the extrapolation to `p_k` matching the printed `a_{k2}` at a relative `1.8 × 10⁻⁵`; two measurements, no embedded producer, and nothing in the row's verdict rests on the constant rather than on the `1/ln z` shape, which is Mertens. `history/staging/recon-0829-farfields2.md` §3e, §5, §6, §7 | | 22 | two-parameter quadratic sieve / Barban–Vehov mean square | Graham estimate, integer case: Chen An [2206.10104v1](https://arxiv.org/html/2206.10104v1), (1.1) and Theorem 1.1, primary statement read | logarithmically averaged full corner coefficient after Mobius inversion | unsigned full-coefficient energy before signed correlation | EXACT-IDENTITY for the smoothed cofactor weight | CLEAN for the auxiliary norm | DERIVED bound, not a twin margin | analytic derivation and bounded algebra check; not a timed census | **LANDED 2026-09-06**: [corner-coefficient-energy.md](corner-coefficient-energy.md) gives O_eta(x log x) squared norms and product bound, retaining all branches; short-endpoint range repaired explicitly. Signed saving and complement OPEN; row 23 prices the sharp transition norm. The easy quadratic-main-term O(D2^2) remainder fails at these long cutoffs. Imported proof not independently re-proved. | | 23 | sharp truncated Mobius divisor sums | de la Breteche–Dress–Tenenbaum [author PDF](https://tenenb.perso.math.cnrs.fr/PPP/Sxz.pdf), (1.5), Theorem 1.1; full source read | full sharp corner after inversion | long-cutoff finite mean square | EXACT-IDENTITY for the coefficient | CLEAN for one-point norms | DERIVED bounds, not a signed margin | analytic derivation and exact finite identities | **LANDED 2026-09-06**: [sharp-corner-transition.md](sharp-corner-transition.md) prices sharp norms and refutes negligible L2 smoothing transfer for sufficiently small fixed eta; signed correlation OPEN. | +| 25 | Schemmel totients and cyclic reduced-residue configuration counts | Defant, Integers 15 (2015) A18, [publisher PDF](https://math.colgate.edu/~integers/p18/p18.pdf), unnumbered definition/formula pp.1-2; Brown, [arXiv:2311.06873v3](https://arxiv.org/pdf/2311.06873v3) (2023), Theorem 3.1/(8) pp.8-9, (16) p.13, Table 1 p.19; [OEIS A059861](https://oeis.org/A059861), exact shift-2 formula and listed values **[SOURCED]** | finite T29/T31/T37 cardinality in accepted return 162 by @zemaj: S2(z#/2), not S2(z#); unique odd CRT lift identifies shift-2 starts with consecutive coprime residues on the odd modulus | none on the exponent or maximum gap | EXACT-IDENTITY at finite cardinality after explicit bijection | CLEAN | PUBLISHED-ANCHOR only | source/translation judgment, scientific CPU 0; no census | **OWNED, attribution-only 2026-09-14**: all three counts are A059861(10,11,12); Brown table directly prints only T29 (v1 p.25, v3 p.19). Served verifier already credits A059861/Schemmel. This adds the precise source map, not novelty or new execution verification; return 162 acceptance unchanged. Every old byte/count preserved; pending audit479 row24 not incorporated. | **Counts (recounted 2026-08-28; the seventeen original rows only, since rows 20 and 21 were added 2026-08-29 at recon grade and are not counted as landings; their own cells are STRONG-ANALOGY/SPLIT and EXACT-IDENTITY/CLEAN).** Seventeen rows, none UNTRIED: fourteen resolved by 2026-08-21, rows 15 and 17 landed 2026-08-28, and row 16 is STALE,