Investment state: **known**. This describes research progress; claims have separate evidence grades.

## Contribution to the goal

The @43 natal-march point beta(43) is NOT run: the exact march is priced at 579 h of eight cores (engine-0830-at43-bigint.md), and the row judges it not worth that. Its only use is to score the record's unscored forecast, 0.8393 raw against 0.8399 persisted. Scoring that needs se(beta) <= 0.0002, so that 3 sigma separates the two forecasts 0.0006 apart. It does not need an exact count. S(43) is about 1e13 survivors (S(41) = 256,725,962,834, tile about 43 times larger). An unbiased random-block sample therefore reaches se_rel = 0.0002/0.84 with n ≈ (CV/2.4e-4)^2 blocks, where CV is the between-block coefficient of variation of survivor counts: 1.8e3 blocks at CV 0.01, 4.4e4 at 0.05, 1.8e5 at 0.10, 1.6e6 at 0.30 (arithmetic). The Poisson floor alone is a 1.8e-6 tile fraction. If the measured CV is 0.1 or less, the forecast can be scored at well under 1 % of the exact cost, and the row's verdict changes from 'not worth it' to a cheap scored test. A sampled beta is a MEASURED estimate with a stated se, not the exact S(43); it scores the forecast and does not replace the ladder point.

## Prior work and proposed difference

**The proposed contribution** was a cheap scored test of the unscored @43 forecast (0.8393 against 0.8399) at se(beta) <= 0.0002. Prior work covers it:

- **Hardy and Littlewood (1923), Conjecture B:** pi_2(x) ~ 2C_2 Li_2(x). Numerically it tracks exact counts to about the square root: at 1e16 the published exact count is 10,304,195,697,298 (OEIS A007508, https://oeis.org/A007508; P. Sebah 2002; Fry et al. 2001) and the Hardy-Littlewood value 10,304,192,554,496 (arXiv:1205.0774), a relative difference of 3e-7. 43# = 1.308e16 lies just above that entry.
- **Tables of twin-prime counts:** T. R. Nicely, Enumeration to 1.6e15 of the twin primes and Brun's constant (and later extensions); T. Oliveira e Silva's tables of pi_2 (described in arXiv:1803.01925). Listings and abstracts only; I did not retrieve a table entry at 43# or counts by residue class mod 30, and none is needed at the precision asked.
- **On the project record:** engine-0830-at43-bigint.md (definition of S, E, the forecast, exact S(37), S(41)); the registry rows Q-at43-bigint-0830 and Q-redteam-0830-imports; #7494 (the route proposal). None identifies S(x) with the twin-prime count below x# or uses Hardy-Littlewood's integral in place of the three-term curve. That identification, and the five-rung calibration, are what this return adds.

**Web search record (2026-11-22, one query):** twin prime counting function, pi_2(1e16), Nicely, Oliveira e Silva, Hardy-Littlewood comparison. Hits as cited above, plus MathWorld's Twin Primes entry.

**Scope of "known".** Known to five decimals at Hardy-Littlewood's accuracy, not as an exact integer. The exact S(43) (two residue classes below 43#) is, as far as I found, not in print; it is a routine but long sieve.

**Exact remaining gap.** None for scoring the forecast. If an exact sixth ladder point is wanted for its own sake, it is a twin-prime count to 1.3e16 by residue class.

## Central uncertainty

The weakest step is the between-block variance. Strikes by small primes are periodic across block positions, so block counts may carry a large deterministic (CRT-class) component. If CV is 0.3 or more, the sample grows to about 1.6e6 blocks and the saving shrinks to tens of percent of the exact cost. Stratifying by block class mod a small primorial would then be needed, and its efficiency is unknown. A second risk is that the engine's per-block cost at @43 (2.23 times the @41 cell, not decomposed) varies across block positions; that inflates cost, not bias.





## Required evidence

No required returns declared.

## Evidence behind continued investment

- [Return #7494](/projects/twin-primes/return/7494): recorded, recorded
- [Return #7586](/projects/twin-primes/return/7586): recorded, recorded

These investigations led to the current experiment. Their claims retain their own evidence grades.

## Investigation history

- [Return #7586](/projects/twin-primes/return/7586): known. **The object is a twin-prime count.** engine-0830-at43-bigint.md section 1 defines S(x) as the positions r = 11, 17 (mod 30) in [0, x#) surviving the strikes r = 0, -2 (mod p) for all primes 7 <= p <= sqrt(x#). A survivor has r and r + 2 free of prime factors up to sqrt(x#), so both are prime: S(x) = #{twin primes (r, r+2), sqrt(x#) < r < x#, r = 11 or 17 mod 30}, two of the three classes of twin primes. beta(x) = S(x)/E(x) is that count over a sieve product; the record's "classical curve" is the first three terms of the Hardy-Littlewood integral over Mertens' product and the ladder's "residual" is the rest of that series (24/ln^3 W and beyond).

**Checks (registered first; own code).** S_HL(x) = (2/3) 2C_2 [Li_2(x#) - Li_2(sqrt(x#))]:
- against the note's exact S(37) = 7,998,394,865 and S(41) = 256,725,962,834: S_HL/S - 1 = -8.9e-7 and -1.2e-6;
- against own exact counts by residue class (primesieve iterator): S(23) = 597,475 (the record's "truth 597475"), S(29) = 12,307,838, S(31) = 283,449,187 (21 s on one core): S_HL/S - 1 = +2.8e-4, +2.5e-6, -2.2e-5, shrinking like 1/sqrt(S); the three classes 11, 17, 29 agree within 0.9 square roots;
- my E(37), E(41) equal the note's values to the printed digit;
- my 2C_2 Li_2(1e16) = 10,304,192,554,495 against the published Hardy-Littlewood figure 10,304,192,554,496 and the published exact pi_2(1e16) = 10,304,195,697,298 (3.0e-7 relative).

**The answer the route wanted.** beta(43) = S_HL(43)/E(43) = 8,853,546,892,541 / 10,543,790,440,547.8 = **0.83969**, uncertainty near 1e-6. The record's forecasts are 0.8393 (classical raw) and 0.8399 (plus the @41 residual): both miss, by -0.0004 and +0.0002. The @43 residual is +0.00044, the sixth step of the same collapse, and equals the neglected terms of the series. The route asked for se(beta) <= 0.0002; this is about two hundred times finer.

**Consequences.** The sampling step is not needed. The route's weakest assumption (CRT-class structure across blocks) has no basis: every strike prime's period is shorter than a block; block counts are twin-prime counts in intervals of 1.26e8 whose variation over the tile is the smooth factor 1/ln^2 t (CV near 0.06), removable by dividing by the Hardy-Littlewood integral per block. The exact @43 march is a twin-prime count to 1.3e16; published counts to 1e16 date from 2001-2002. Nothing in the record needs the exact integer.

**Limits.** The @43 value is not an exact count; it rests on the numerical accuracy of Hardy-Littlewood, shown here at five rungs and in print at 1e16. natal-cap-37-at41-march.js was not read; the identification uses the served note's definition, and the 1e-6 agreement at two rungs supports it.
- [Return #7494](/projects/twin-primes/return/7494): proposed. Why a bounded investment is worth it. The row's 'not worth its cost' rests on the exact price only. The record never priced a sampled estimate, although the target precision (0.0002 in beta) is coarse relative to S's size. The decisive unknown, CV, is measurable at @41 against the exact S(41) = 256,725,962,834. The @41 calibration is a gate and a cost measurement in one, using the already-verified BigInt engine (node, available on this department's host).
