Exhaustive computation of prime occurrence on the 6N±1 skeleton up to 10¹0 (all 1. 67×10⁹ centres scanned; 2. 74×10⁷ twin pairs identified, shells S1–S10), together with a factor-resolved local model for the conditional density of twinprimes. Conditioning on the factorisation of the centre N, the conditional twindensity is shown to be proportional to E (N) = ∏ₐ|₍, ₐ>₃ (q−1) / (q−3). The modelis derived within the Hardy–Littlewood heuristic, validated factor by factoragainst ~10⁸ nodes (observed/predicted = 1. 000 ± 0. 008 across strata spanning a3. 3× enrichment range, selecting it over the naïve ∏ q/ (q−2) ), and shown toreproduce the squarefree bias reported by Puszkarz (2018, arXiv: 1807. 00406) towithin −0. 2% at the matching 10¹0 scale. This deposit contains: the five result tables (S1–S10, complete factorisation), the computation and figure-generation code, the two figures, and the manuscript (LaTeX + PDF). Scope. This is experimental/computational number theory. The model is avalidated Hardy–Littlewood heuristic, not an unconditional theorem; nothing herebears on the infinitude of twin primes. Priority for the underlying phenomenonbelongs to W. Puszkarz (2018) ; the cumulative twin counts coincide with OEISA007508.
Ruqing Chen (Sun,) studied this question.