Skip to content

Documentary state: DRAFT_UNACCEPTED. This draft has NOT been editorially admitted. Mathematical claims require independent checking.

OAI-26 family 012: largest prime factors of consecutive integers

Editorial status: DRAFT_UNACCEPTED. NO_CERTIFICATION. The article documents a publisher-reported theorem, not an independently verified proof.

The question in plain language

Every positive integer factors into primes. Its largest prime factor records the largest prime appearing in that factorization. Two consecutive integers have no common prime divisor, but this does not automatically mean their largest prime factors behave independently.

The Erdős–Pomerance joint Dickman conjecture concerns the sizes of the largest prime factors of consecutive integers. The pinned OAI-26 overview claims that a September 24, 2026 manuscript establishes their asymptotic independence in ordinary natural density. It also claims that exactly half of positive integers, in the density limit, have a smaller largest prime factor than their successor.

A small example

The integers 14 and 15 are consecutive. Their largest prime factors are 7 and 5, respectively. For 20 and 21, they are 5 and 7. Individual examples fluctuate, and counting a modest initial interval cannot establish an asymptotic law.

Write the largest prime factor of an integer \(n>1\) as \(P^{+}(n)\). A natural normalized size is

\[ \frac{\log P^{+}(n)}{\log n}. \]

The Dickman distribution describes the limiting behavior of smooth numbers, integers whose prime factors are all relatively small. A joint Dickman law for \(n\) and \(n+1\) asks whether the two normalized measurements behave, asymptotically, like two independent samples from that limiting law.

Independence here means a factorization of limiting joint probabilities, not that neighboring integers are arithmetically unrelated.

What is known

The pinned Family 012 catalogue identifies one manuscript, The joint Dickman law for consecutive integers (September 24, 2026). The publisher's source overview attributes two conclusions to it:

  • The logarithmic sizes of the largest prime factors of \(n\) and \(n+1\) converge jointly to independent Dickman-type limits, when integers are counted with ordinary natural density.
  • The set of integers satisfying \(P^{+}(n)<P^{+}(n+1)\) has natural density one half.

The distinction between ordinary density and logarithmic averages is material. A logarithmically weighted average can be easier to control and is not interchangeable with an ordinary count divided by \(X\).

The comparison probability of one half is consistent with an independent, identically distributed continuous limiting pair, where equality has limiting probability zero. However, transferring weak convergence to this strict inequality requires checking that the limiting distribution has no mass on the diagonal and that the normalizations and finite-sample comparison align.

These are the indexed manuscript claims. The exact theorem statements, quantified convergence, treatment of \(n=1\), and any effective error terms require direct examination of the pinned PDF.

GCL verification state

The source snapshot is pinned to openai/math@1581d4cbdd69a0faf34b2089614fe2a814316a6a. The Family 012 catalogue has one manuscript path, and the overview states the joint-law and one-half comparison claims.

This pass has not independently checked the paper's analytic-number-theoretic proof, its treatment of correlations between neighboring integers, or any specific sieve or distributional estimates. It establishes no Family 012 paper-to-Lean mapping, Comparator acceptance, semantic audit, or MATHCERT certification.

The documentary artifact is consequently marked DRAFT_UNACCEPTED and NO_CERTIFICATION.

What remains

A source critic should inspect the pinned PDF and record the precise joint convergence theorem, normalizing variables, test sets, and conditions for ordinary density. The audit should identify where independence is proved rather than merely assuming it from coprimality, and track any uniformity needed to infer the strict largest-prime-factor comparison.

Review must separately check the distribution's continuity on the diagonal, any negligible exceptional set, and whether equality of neighboring largest prime factors is excluded or shown to have density zero.

After source criticism, validate mathematical Markdown, inspect rendered HTML, record a content-bound editorial disposition, and obtain protected merge/readback. These are documentary controls, not mathematical proof certification.

Sources and claim boundaries

Boundary: NO_CERTIFICATION. This explains a reported resolution of a joint-distribution conjecture and does not independently establish the result or certify a formal proof.


Indexed primary manuscript paths: