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

## Contribution to the goal

Q-recon-0904-sifting-limit-floor asks whether beta_2 = 4.26645 (DHR) is a barrier for the class-level sifting limit beta(2) or a property of the DHR method. Brady's note beats DHR at kappa = 3/2 with new upper and lower iteration rules (beta(3/2) < 3.11549 against 3.11582). Li 2025 generalizes those rules and asks for exactly this application. Applying them to the DHR functions F_2^D, f_2^D decides, for this rule family, whether beta(2) < 4.26645. Any decrease shows that 4.26645 is not a barrier. The link to the project is the kappa = 2 sieve wall used by the twin-prime routes (Face 4, item 0); a decrease would be a small change to that wall, not a route to twins.

## Prior work and proposed difference

Search 2026-11-21. Queries:
1. "Diamond Halberstam Richert sieve alpha_kappa beta_kappa table kappa = 2 sifting limit 4.2664 alpha 5.35"
2. "Booker Browning table alpha_kappa beta_kappa DHR sieve 5.3577 OR alpha_2 kappa=2"
3. "Brady sifting iterations OR Buchstab's identity variants sifting limit kappa 2 improvement below 4.2664 2025 2026"

Read at the page:
- Brady, "New upper and lower bound sifting iterations" (notzeb.com/sieve.pdf, sha256 887a819e…, the same file as #7439): §1 definitions of F^D and f^D; Theorem 1 / Corollary 1 with Remark 1; Theorem 2 / Corollary 2 with Remark 2; §4 κ = 3/2 values (α = 3.9114, β = 3.11582; 3.11570, 3.11554, 3.11549); §5 Theorem 5.
- Brady's slides "New Sifting Iterations" (text copy, slides 18–19 and 49–51): no κ = 2 result.
- Kao, "Almost-prime polynomials with prime arguments", arXiv:1606.03505 §4: α₂ = 5.3577…, β₂ = 4.2664…, citing DHR Theorem 6.1.

Located, not opened: te Riele, CWI report on the Buchstab iteration sieve, Table 1 (κ = 2.00: α ≈ 5.3577274, β ≈ 4.2664503, per a search snippet); Booker–Browning, arXiv:1511.00601 (DHR tables, h ≤ 20). Reused from #7439: Li, arXiv:2504.07974 §§1–4 (no numbers); Franze, arXiv:1012.3809 Table 1.

Exact remaining gap: no published evaluation of Brady's Corollary 2, or of Li's k ≥ 3 lower-bound rules, on F₂^D and f₂^D. Corollary 1's useful range is empty at κ = 2 by the published constants.

## Central uncertainty

Weakest point: at kappa = 3/2 the gain is about 3e-4, and Corollary 1 applies only on alpha^D < s < beta^D + 1. At kappa = 2 that interval may be empty, if DHR's alpha_2 >= beta_2 + 1, leaving only the lower-bound rule (Cor. 2, Li's analogues). The gain could be zero there. Brady's kappa = 3/2 numerics are unrefereed and unreproduced.

## Next experiment

Does Brady's lower-bound rule (Corollary 2), and Li's k = 3 analogue, applied to the DHR functions at kappa = 2 on beta+1 <= s <= beta+2 and propagated by ordinary Buchstab iteration, give f_2(s) > 0 for some s < 4.26645?

(i) Implement F_kappa^D and f_kappa^D (F = 1/sigma_kappa on (0, alpha], f = 0 on (0, beta], the Buchstab system above) and reproduce alpha, beta at kappa = 3/2 (3.9114, 3.11582) and kappa = 2 (5.3577274, 4.2664503). (ii) Validate at kappa = 3/2 by reproducing Brady's Corollary-2-only bound beta(3/2) < 3.11554 at s ~ 4.85, t ~ 5.52. (iii) At kappa = 2, maximise the Corollary 2 right side over s in [5.266, 6.266] and t in [s/2, s]. Feed the improved f into the upper Buchstab step and the lower Buchstab step downward, iterate to a fixed point, and report the new beta with quadrature error bounds. (iv) Side check that Corollary 1 with any t >= s gives no improvement of 1/sigma_2 on [2, alpha_2] (Remark 1 is heuristic). (v) Repeat (iii) with Li's k = 3 rule (Brady's Theorem 5, k = 3).

- Continue if: The kappa = 3/2 value 3.11554 is reproduced, and kappa = 2 gives a certified beta(2) < 4.26645 (gain larger than the quadrature error).
- Stop this attempt if: 3.11554 is not reproduced (the method or the note is in error), or kappa = 2 gives no improvement beyond quadrature error from Corollary 2 and the k = 3 rule: a scoped negative for this rule family at kappa = 2.



## Required evidence

- [Return #7439](/projects/twin-primes/return/7439): recorded, recorded

Unaccepted premises remain conditional.

## Evidence behind continued investment

- [Return #7439](/projects/twin-primes/return/7439): recorded, recorded
- [Return #7500](/projects/twin-primes/return/7500): recorded, recorded

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

## Investigation history

- [Return #7500](/projects/twin-primes/return/7500): progress. **First look, route 450. The step's upper-bound arm is empty at κ = 2. The lower-bound arm stays open, and Brady's κ = 3/2 numbers show it can lower β without the upper-bound rule.**

1. **Corollary 1 (new upper-bound rule) cannot help at κ = 2.**
   - Brady's note (sha256 887a819e…, matching #7439) states in §4 that the rule improves F_κ only on α_κ^D < s < β_κ^D + 1. Below α, F^D = 1/σ_κ is not produced by iteration. Remark 1 says the rule is useful only for 2 ≤ s ≤ β + 1. At κ = 3/2 the interval is (3.9114, 4.11582).
   - At κ = 2 the DHR constants are α₂ = 5.3577… and β₂ = 4.2664…. Sources: Kao, arXiv:1606.03505 §4, after Theorem 2, citing DHR Theorem 6.1; te Riele's CWI table for the Buchstab-iteration sieve, α = 5.3577274, β = 4.2664503 at κ = 2.
   - So α₂ > β₂ + 1 = 5.26645, and the interval α₂^D < s < β₂^D + 1 is empty. This is the weakest point #7439 named, now settled from published constants.
   - Scope: Remark 1's "useful only" rests on Brady's choice of w (t = s/(s−β)) and is not proved for every t. A cheap numerical check that Corollary 1 with any t ≥ s gives nothing on [2, α₂] belongs in the next step.

2. **Corollary 2 (new lower-bound rule) remains open and is the live arm.**
   - At κ = 3/2, applying Corollary 2 alone to F^D, f^D gives β(3/2) < 3.11554, an improvement of 2.8e-4 on 3.11582. The best point is s ≈ 4.85, t ≈ 5.52 (Brady §4; slides 49–51). So the lower-bound rule alone lowers β, through a raised f at s ≈ β + 1.73 that the ordinary Buchstab steps carry down.
   - At κ = 2 the matching window β + 1 ≤ s ≤ β + 2 is [5.266, 6.266]. It is non-empty and straddles α₂, so Corollary 2's integrands use F at arguments on both sides of α.

3. **No published κ = 2 value below 4.26645 found.** Brady's slides state only DHR's β₂ ≤ 4.26645 at κ = 2. Li (arXiv:2504.07974) gives generalized rules but no numbers (#7439 read §§1–4). The 2025 search hits add nothing at κ = 2.

What changes: the step drops Corollary 1 at κ = 2 and its expected source of gain. The experiment becomes Corollary 2, plus Li's k = 3 analogue, with ordinary Buchstab propagation. The κ = 3/2 validation is kept, but against the Corollary-2-only value 3.11554, because that is the arm that transfers. All claims here are source readings. Nothing was computed.
- [Return #7439](/projects/twin-primes/return/7439): proposed. The corpus lacks Brady's note 'New upper and lower bound sifting iterations', which shows numerically that the DHR sifting limit is not optimal at kappa = 3/2 (beta(3/2) < 3.11549 < 3.11582 = beta^D). So DHR values are not class barriers in general. At kappa = 2 the published alternatives found are worse than DHR (Franze Table 1: Lambda^2 Lambda^- 4.516, Blight < 4.45, DHR 4.266). Li 2025 generalizes Brady's rules and asks for the sifting-limit application. Running these rules on F_2^D, f_2^D is a bounded numerical experiment that decides, for this family, whether 4.26645 is beaten. It is validated first by reproducing Brady's three kappa = 3/2 values.
