// ============================================================================ // THE FULL A144311 LADDER, ADOPTED AS TRUSTED DATA — 22 TERMS TO x = 79 // // Question: what do the eight terms above our exact ladder (x = 47..79, // OEIS A144311, trusted under the series rule of 2026-08-20) settle for free — // the x = 37 outlier, the direction of G2(x#)/x^2, and the primorial-base // threshold column S(x) = ln(x^2/G2)? // // Provenance of the trusted series, read at the OEIS entry 2026-08-20: // a(1)-a(7) Andrew Carter, Sep 17 2008 // a(8)-a(16) Max Alekseyev, Nov 18 2009 // a(17)-a(22) Jinyuan Wang, Nov 26 2024 — derivation PUBLIC, a C++ // branch-and-bound DFS at https://oeis.org/A144311/a144311.cpp.txt // (plist starts at 5: the program works inside the 6-wheel, and // its pskip encodes the -2 class; entry comment a(n) == 5 mod 6 // for n > 1 is checked below as a transcription guard). // Convention: A144311 counts consecutive integers each hitting {1,-1} mod some // prime, which is our G2 - 1 (GLOSSARY: "the sequence is A144311 + 1"). // // Honest doubt: terms 15-22 are single-witness (no independent computation in // print; our enumeration ends at 43#). The series rule trusts them during // research; in-house verification is a PAPER-PHASE deliverable (TODO 1c). // The 14-term overlap with our custody ladder is checked exactly below — a // disagreement there would void the trust, so it is a hard assert. // // Everything integer is BigInt; logs are doubles over small-integer factors // (relative error ~1e-15, far under the 4 significant figures quoted). // Runtime: instant. // ============================================================================ 'use strict'; // A144311 verbatim (a-values). Index n = 1..22, prime = n-th prime. const A = [1, 5, 11, 29, 41, 65, 107, 149, 203, 257, 347, 527, 545, 617, 707, 869, 965, 1079, 1283, 1397, 1529, 1709]; const P = [2, 3, 5, 7, 11, 13, 17, 19, 23, 29, 31, 37, 41, 43, 47, 53, 59, 61, 67, 71, 73, 79]; // Our exact custody ladder (research/exact-g2-ladder.js), G2 convention: const CUSTODY = [2, 6, 12, 30, 42, 66, 108, 150, 204, 258, 348, 528, 546, 618]; // --- transcription guards -------------------------------------------------- for (let n = 2; n <= 22; n++) { if (A[n - 1] % 6 !== 5) throw new Error(`a(${n}) = ${A[n - 1]} not 5 mod 6`); } let overlapOK = true; for (let i = 0; i < CUSTODY.length; i++) { if (A[i] + 1 !== CUSTODY[i]) { overlapOK = false; break; } } if (!overlapOK) throw new Error('custody overlap FAILED — trust voided'); // --- exact ladder quantities ------------------------------------------------ // W = x#, D = prod_{2
s + Math.log(f), 0); } const rows = []; for (let n = 0; n < 22; n++) { const x = P[n], g2 = A[n] + 1; const primes = P.slice(0, n + 1); const lnW = lnBig(primes); // theta(x) const lnD = lnBig(primes.filter(p => p > 2).map(p => p - 2)); const lnm = lnW - lnD; // m = W/D const m = Math.exp(lnm); const c2p = x >= 11 ? g2 / (m * lnD) : null; // c2' = G2/(m lnD), x>=11 const S = Math.log(x * x / g2); // S(x) = ln(x^2/G2) rows.push({ n: n + 1, x, g2, m, lnD, c2p, S, gxx: g2 / (x * x) }); } // --- output ----------------------------------------------------------------- console.log('THE FULL A144311 LADDER IN G2 CONVENTION (22 terms, trusted)'); console.log('custody overlap x=2..43: EXACT, 14/14 terms | a(n)=5 mod 6: 21/21'); console.log(''); console.log(' n x G2 m lnD c2\' G2/x^2 S(x)=ln(x^2/G2)'); for (const r of rows) { console.log( String(r.n).padStart(3) + ' ' + String(r.x).padStart(2) + ' ' + String(r.g2).padStart(5) + ' ' + r.m.toFixed(2).padStart(7) + ' ' + r.lnD.toFixed(3).padStart(7) + ' ' + (r.c2p === null ? ' --' : r.c2p.toFixed(4).padStart(7)) + ' ' + r.gxx.toFixed(4).padStart(8) + ' ' + r.S.toFixed(4).padStart(8)); } console.log(''); const band = rows.filter(r => r.x >= 11 && r.x <= 31).map(r => r.c2p); console.log(`c2' band x=11..31 (custody): [${Math.min(...band).toFixed(4)}, ${Math.max(...band).toFixed(4)}]`); console.log(`c2'(37) = ${rows[11].c2p.toFixed(4)} (the outlier under test)`); const after = rows.filter(r => r.x >= 41).map(r => `${r.x}:${r.c2p.toFixed(4)}`); console.log('c2\' x=41..79: ' + after.join(' ')); const above = rows.filter(r => r.x >= 41 && r.c2p > rows[11].c2p).length; console.log(`terms x>=41 above c2'(37): ${above} of ${rows.filter(r => r.x >= 41).length}`); console.log(''); const Smax = rows.reduce((a, b) => (b.S > a.S ? b : a)); console.log(`S(x) max over the 22 primorial bases: ${Smax.S.toFixed(4)} at x = ${Smax.x}`); const s47 = rows.filter(r => r.x >= 47).map(r => r.S.toFixed(4)).join(' '); console.log(`S(x) x=47..79: ${s47}`); const g11 = rows.find(r => r.x === 11).gxx, g79 = rows.find(r => r.x === 79).gxx; console.log(`G2/x^2: ${g11.toFixed(4)} at x=11 -> ${g79.toFixed(4)} at x=79`); let desc = 0, tot = 0; for (let i = rows.findIndex(r => r.x === 11); i < rows.length - 1; i++) { tot++; if (rows[i + 1].gxx < rows[i].gxx) desc++; } console.log(`descents in G2/x^2 over steps from x=11: ${desc} of ${tot}`); // ============================================================================ // OUTPUT — EMBEDDED, do not hand-edit. Regenerate: // node research/qc/embed.js research/a144311-full-ladder.js // invocation: node research/a144311-full-ladder.js // code-sha256: 1c3d6a80160ce3bb6c81b3bc957884723cd0630d0d56bf6b04f2a65ccbe75232 // out-sha256: a658deda243b350b376dd994d4a8426b22c0c8da364eb4be800aa3ccdbd47880 // streams: stdout // node: v22.21.0 // embedded: 2026-08-20 // elapsed: 0.1 s // ============================================================================ // THE FULL A144311 LADDER IN G2 CONVENTION (22 terms, trusted) // custody overlap x=2..43: EXACT, 14/14 terms | a(n)=5 mod 6: 21/21 // // n x G2 m lnD c2' G2/x^2 S(x)=ln(x^2/G2) // 1 2 2 2.00 0.000 -- 0.5000 0.6931 // 2 3 6 6.00 0.000 -- 0.6667 0.4055 // 3 5 12 10.00 1.099 -- 0.4800 0.7340 // 4 7 30 14.00 2.708 -- 0.6122 0.4906 // 5 11 42 17.11 4.905 0.5004 0.3471 1.0581 // 6 13 66 20.22 7.303 0.4469 0.3905 0.9402 // 7 17 108 22.92 10.011 0.4707 0.3737 0.9843 // 8 19 150 25.61 12.844 0.4559 0.4155 0.8782 // 9 23 204 28.05 15.889 0.4577 0.3856 0.9529 // 10 29 258 30.13 19.185 0.4463 0.3068 1.1816 // 11 31 348 32.21 22.552 0.4791 0.3621 1.0158 // 12 37 528 34.05 26.107 0.5939 0.3857 0.9527 // 13 41 546 35.80 29.771 0.5123 0.3248 1.1245 // 14 43 618 37.54 33.485 0.4916 0.3342 1.0959 // 15 47 708 39.21 37.291 0.4842 0.3205 1.1379 // 16 53 870 40.75 41.223 0.5179 0.3097 1.1721 // 17 59 966 42.18 45.266 0.5059 0.2775 1.2819 // 18 61 1080 43.61 49.344 0.5019 0.2902 1.2370 // 19 67 1284 44.95 53.518 0.5337 0.2860 1.2516 // 20 71 1398 46.25 57.752 0.5233 0.2773 1.2826 // 21 73 1530 47.56 62.015 0.5188 0.2871 1.2479 // 22 79 1710 48.79 66.359 0.5281 0.2740 1.2946 // // c2' band x=11..31 (custody): [0.4463, 0.5004] // c2'(37) = 0.5939 (the outlier under test) // c2' x=41..79: 41:0.5123 43:0.4916 47:0.4842 53:0.5179 59:0.5059 61:0.5019 67:0.5337 71:0.5233 73:0.5188 79:0.5281 // terms x>=41 above c2'(37): 0 of 10 // // S(x) max over the 22 primorial bases: 1.2946 at x = 79 // S(x) x=47..79: 1.1379 1.1721 1.2819 1.2370 1.2516 1.2826 1.2479 1.2946 // G2/x^2: 0.3471 at x=11 -> 0.2740 at x=79 // descents in G2/x^2 over steps from x=11: 10 of 17 // ============================================================================ // READINGS // // 1. THE TRUST'S HARD GATE PASSED. The 14-term overlap with our exact custody // ladder (x = 2..43) matches EXACTLY, and all 21 applicable terms satisfy // the entry's own a(n) = 5 mod 6 invariant. Had any overlap term differed // the script throws and nothing below exists. Two independent methods // (Wang's branch-and-bound vs our tile-major enumeration) agreeing on 14 // terms is the calibration that makes trusting the other 8 rational. // 2. THE x = 37 OUTLIER IS CONFIRMED AN OUTLIER — by all ten trusted terms // above it. c2'(37) = 0.5939 and NOT ONE of x = 41..79 reaches it (max // 0.5337 at x = 67). But the flat-band reading dies with the same table: // the x >= 41 terms run [0.4842, 0.5337], ABOVE the custody band // [0.4463, 0.5004] of x = 11..31. So c2' drifts upward with x and 37 is a // single spike on that drift, not a step to a new level. The 1c question // ("was 37 an outlier?") is settled YES for free; what replaces it is // "why does c2' drift?", which is a different question. // 3. G2(x#)/x^2 LEANS FALLING: 0.3471 (x = 11) -> 0.2740 (x = 79), with 10 of // 17 steps descending. This is a drift over 22 exact-or-trusted terms, not // a fit and not a law; 1d's central question stays open but the trusted // tail leans the same way the custody head did. // 4. THE THRESHOLD COLUMN TIGHTENS. S(x) = ln(x^2/G2) has max 1.2946 at // x = 79 over all 22 primorial bases, against 1.1816 (x = 29) on custody // terms alone — every trusted term raised the reachable max, and the tail // is still rising. (The integer-base threshold of item 1d is a separate // object and is not recomputed here.) // 5. LIMITS. Terms 15-22 are single-witness (Alekseyev 2009 for 8-16, Wang // 2024 for 17-22, derivation public in the entry's C++). Trusted under the // series rule of 2026-08-20; in-house verification is a paper-phase // deliverable (TODO 1c). No fits were performed here; exponent-control.md // owns the fitting question and its 22-term refit is queued separately. // ============================================================================