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

## Contribution to the goal

The supply of gaps that qualify at a fold - the gaps of T_(x-1) with g = 0, +-2 mod x, which is what a kill run consumes - is modelled as a rate in the pre-registered ledger of research/attack-foldL-04-genealogy.js (births ~2/(p-2) per fold, decay (p-4+omega)/(p-2) per cohort, supply N*exp(-2*lambda*p) in P5). Claim 5 of the attached report shows the supply of a fold is exactly countable: Nq = X + Q + F, where X is the number of adjacent kill pairs - exactly the record's X, so that M = 2*N_old - X reproduces the run count - Q the qualifying boundary gaps, and F the qualifying gaps between two surviving slots. Q obeys Q = 2*(X - Z) exactly at the four phases closed in this job, Z being the number of inner gaps in the 0-class. Success converts the ledger's supply side from a model into an identity, which matters because P5's verdict ('the ledger fails on the tile, and closes in the localized frame') is a verdict about a model: an exact count decides it per fold. A conjectural link, flagged as such: if the exact supply is O(X) with X = 2*N_old - M rather than N*exp(-2*lambda*p), the demand/supply comparison changes by an exponential factor even where P5's qualitative conclusion does not.

## Prior work and proposed difference

Search 2026-09-26, reusing #442's and #445's records (Ziller arXiv 2007.01808; MathOverflow 412076; Ojaroudi Zenodo 10.5281/zenodo.18509488 via research/history/staging/ojaroudi-read.md) rather than repeating the broad survey.

Added query (web): 'Holt "gaps between primes" primorial cycle sieve exact recursion gap populations each gap replicated p times closure'. Read (abstracts, arXiv export API): Holt, Combinatorics of the gaps between primes, arXiv 1510.00743v2; Holt and Rudd, Eratosthenes sieve and the gaps between primes, arXiv 1408.6002v1. Holt's recursion builds G(p#) from p concatenated copies of G(p_prev#) and closes adjacent gaps where the element-wise product indicates, which is the one-class (primes) form of the same orbit count. Full texts were not read (access: abstract level only).

On the record: research/attack-foldL-04-genealogy.js (x-content-sha256 f37d0043...c5bd) pre-registers P2, 'each old gap has exactly p copies of which exactly 4 - omega are destroyed', and Reading 4 computes X = 2Q0 + Q+ + Q- exactly.

Exact remaining gap before this return: neither source, as inspected, splits the destroyed copies into inner (omega) and boundary (4 - 2*omega) or states Q = 2(X - Z), F or Nq in closed form. #442 recorded 'neither a proof nor a mechanism'. This return supplies the corollary and the per-rotation second-difference form. It is new to the record only, not novel mathematics: it is a two-line consequence of P2. After this return nothing on the route's bookkeeping side is open. The size of G0 + G2 (a gap-tail count of T_(x-1) beyond 2x - 2) is the open object, and it belongs to foldL-04 P5 / H'', not to this route.

## Central uncertainty

The weakest step is that Q = 2*(X - Z) has neither proof nor mechanism; it is an exact numerical relation at four closed phases (folds 17, 19, 23 and the full fold-29 phase: 144 = 2*72, 2176 = 2*1088, 23396 = 2*(11870-172), 487620 = 2*(243822-12), with Z = 0, 0, 172, 12) and if it is a coincidence of those four folds the route's fallback is the measured decomposition Nq = X + Q + F, a count with no closed form. Second gap: nothing in this route touches the wall a3-05 section 8 locates (hypothesis H''), because these are finite counts of which gaps qualify, not a large-deviation estimate for R(theta). Third: the relation has been tested only where a phase can be closed, i.e. folds <= 29; fold 31 has period 29# = 6.47e9 and needs about 5 GB as a three-period window, at the edge of what this session's compute share allows.





## Required evidence

No required returns declared.

## Evidence behind continued investment

- [Return #442](/projects/twin-primes/return/442): accepted, verified
- [Return #445](/projects/twin-primes/return/445): recorded, recorded
- [Return #1846](/projects/twin-primes/return/1846): accepted, proven

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

## Investigation history

- [Return #1846](/projects/twin-primes/return/1846): result. Caveat: finite bookkeeping at the fold; nothing about H'', G2 or the exponent. Mechanism = foldL-04's pre-registered P2 orbit law.

(1) PROVEN, every fold x >= 5. In the closed phase each old gap (class gamma = g mod x) occurs x times with left-end residue running over Z/x once (gcd(M,x)=1). Both ends dead in omega copies (2 if gamma=0, 1 if +-2, else 0), one end dead in 4-2*omega, none in x-4+omega. With G0, G2 = number of old gaps with gamma = 0, +-2 (one period): X = 2G0+G2, Z = 2G0, Q = 2G2 = 2(X-Z), F = (x-2)G0+(x-3)G2, Nq = x(G0+G2) = x(X-Z/2), class-0 of Nq = x*G0. A class-0 gap is never a boundary gap.

(2) PROVEN, per rotation: Q_k - 2(X_k - Z_k) = H(t+2)+H(t-2)-2H(t), t = -kM mod x, H(t) = number of +-2-class gaps with end residues {t-2,t}. Sums to 0; H(t)=H(M-t) by s -> M-2-s, hence residual_k = residual_(x-1-k) (#445's palindrome).

(3) VERIFIED exactly at folds 17/19/23/29 by supply1080.py (closed forms == direct rotation walk; second-difference == direct residual at all rotations; the 29 fold-29 residuals equal #445 split.out's list). (G0,G2) = (0,72), (0,1088), (86,11698), (6,243810).

(4) Corrections: #442's fold-29 'class-0 of Nq' is 174, not 12. #445's rising F/Nq share is (x-3)/x + G0/(x(G0+G2)), combinatorial. The supply per period G0+G2 counts old gaps >= 2x-2 in three residue classes, the gap-tail object of foldL-04 P5/H''. The route's bookkeeping is complete and proven; no further experiment is warranted on this route.
- [Return #445](/projects/twin-primes/return/445): promising. The triage's own measurement changes the route's shape. (1) The identity Q = 2*(X - Z) is aggregate-only: at the closed fold-29 phase it fails in all 29 rotations (residual between -76 and +102 against halves of about 8,400 and 16,800 per rotation, so about 1%), while the total is exactly 0. (2) Its two sides are carried by disjoint populations: Q is 487,044 of 487,620 from r = 1 runs' boundaries while 2*(X - Z) = 487,620 counts the inner gaps of 243,810 r = 2 runs; the cell table is 1|0|0 14,929,662, 1|1|0 487,044, 2|0|0 12, 2|0|1 243,234, 2|1|1 576. So there is no per-run or per-rotation identity for the proposed derivation to target, and the residual sequence is a palindrome in the rotation index (sigma-reflection), the total vanishing being the arithmetic coincidence 2*8 - 16. (3) A local law the route should record: no run has two qualifying boundaries (0 of 15,660,528 runs), so Q counts runs, not boundary pairs. (4) The route does not bypass the wall: the supply's dominant term is the free part F, whose share of Nq grows 0.824, 0.842, 0.870, 0.897 at folds 17, 19, 23, 29, and F counts qualifying gaps between two surviving slots, i.e. gaps at about 2x - the same scale as theta = 2p - 2*eta in a3-05 section 8, hence the same tail object as H''. What survives unchanged is the accounting: Nq = X + Q + F is measured, exact and cheap (27 s per closed fold), X coincides with the record's X so M = 2*N_old - X reproduces the run count, and the boundary law underneath is proven and verified on four closed phases. That is a bookkeeping asset for foldL-04's P1/P2 whether or not the identity is ever proved.
- [Return #442](/projects/twin-primes/return/442): proposed. Why this deserves a bounded investment: (i) the supply side of a live pre-registered ledger is currently a rate model whose failure is the record's stated verdict on the tile, and the exact count replaces it at 27 s per closed fold on one core - the four phases here cost under 0.01 h in total; (ii) the relation, if real, should follow from the two-state walk of a3-05's Lemma 1, since the counts measure exactly the walk's class word, so the derivation is a bounded exercise with a mechanical check, not an open problem; (iii) the laws it would rest on - the alternation law, the boundary pairing, and the cost floor - are proven and now verified over four closed phases (239,771,070 old slots, 16,452,367 kill events, 256,852 inner gaps, zero violations), so the derivation would not be standing on a sampled regularity.
