Skip to content

Erdős Problem #369

Editorial depth: Reviewed explanatory article — evidence qualified.
Snapshot classification: proved (not automatically the live result).
Subject areas: number theory.

Read this status correctly

The status above is reproduced from GCL's immutable external-catalogue snapshot. It is not a new claim of current mathematical resolution, a source-semantic review, an independent proof check, or GCL certification. A published proof and a GCL replay are separate events.

The question in plain language

Erdős Problem 369 concerns strings of consecutive integers whose prime factors are small. If \(P^+(m)\) denotes the largest prime factor of \(m\) (with the harmless convention \(P^+(1)=1\)), the literal question recorded by the catalogue is: for every real ε > 0 and integer k ≥ 2, does every sufficiently large integer n contain k consecutive integers in {1, …, n} for which \(P^+(m) ≤ n^ε\)?

The wording matters. It does not require the block to lie near n, and it measures smoothness against n, not necessarily against each member m. For fixed k, the first k positive integers already form such a block once n is large enough. Thus the literal statement is elementary. The serious nearby question is whether the block can be forced into a region such as [n/2, n].

This distinction is present in the live Erdős Problems entry and in the formal source discussed below. It is not a cosmetic reformulation: a theorem about every large n near n is substantially stronger than the literal initial-block observation.

A small example

Here is an exact finite check of the stronger near N shape, included only as an illustration. Take N = 64, k = 3, and ε = 1/2. The consecutive block 48, 49, 50 lies in [32, 64], and

\[ 48 = 2⁴ · 3, 49 = 7², 50 = 2 · 5². \]

Therefore the largest prime factors are 3, 7, and 5. Their squares are 9, 49, and 25, each at most 64, so each largest prime factor is at most √64 = 8. This verifies the finite instance \(P^+(m) ≤ m^{1/2} ≤ 8\) for these three values by exact integer arithmetic.

The check does not establish an asymptotic theorem, does not identify a threshold for all N, and does not validate any Lean implementation. It simply makes the two notions—consecutive and smooth—concrete, while showing that the near N variant is not vacuous at small scale.

What is known

The frozen catalogue evidence is teorth/erdosproblems commit 536408716724a2402bdfa7c1d764bf83d64a8c89, file data/problems.yaml, Git blob 4e60263a80af6207aca494d2e1388e44fa372889. Its entry 369 records informal status proved, formal status Lean, and formalized: yes; the entry dates those fields 27 March 2026 and the formalized field 3 August 2026. Those are catalogue facts, not by themselves a proof or a GCL decision.

The primary published result most directly related to the older literature is Antal Balog and Trevor D. Wooley, “On strings of consecutive integers with no large prime factors,” Journal of the Australian Mathematical Society (Series A) 64 (1998), 266–276, DOI 10.1017/S1446788700001750. Its Theorem 1, p. 267, gives infinitely many long strings of consecutive integers with no prime factor above a small power of the scale. This is an “infinitely many” result, not automatically a block in [N/2, N] for every sufficiently large N.

The accessible primary manuscript for the stronger every-large-N interpretation is Yueer Yang, “Proof of Erdős Problem #369: Consecutive Smooth Integers Near N,” Zenodo record 19361034, DOI 10.5281/zenodo.19361034, created 31 March 2026. The deposited TeX file has Zenodo MD5 1af08fc9065365a70cc405287e7a9b5f and SHA-256 00a1aedb6d5188bf33ae6a162ecac9feb9bcbf6983bc988b7354bcd986d5ebfa. Its Theorem 1 asserts that, for every ε > 0 and k ≥ 2, all sufficiently large N have a block of k consecutive integers in [N/2, N] with \(P^+(m) ≤ m^ε\). The written proof uses small Euler-product ratios, the Chinese remainder theorem, cyclotomic factorization, and a covering argument from powers of 2 and 3. This manuscript is a source of an external mathematical claim; it is not a GCL certification.

The live entry at erdosproblems.com/369, accessed during this audit, labels the problem “PROVED (LEAN)” but also explains that the literal formulation is trivial and that the near N readings are variants. Its discussion cites Balog–Wooley and the Yang construction. The page reports a last edit of 28 April 2026, so its status text is later than the frozen catalogue packet and must be treated as a separate dated revision.

GCL's current state

The issue-specific packet classifies the existing Programme triage as OUTSIDE_STRICT_OPEN_TRIAGE and says that no statement file for problem 369 is present in the admitted FC-GDM-85F86371 provider snapshot. The provider’s exact upstream revision is 85f863718beeec7b58a3a1926ee92e3472bc2020; querying that revision confirms that FormalConjectures/ErdosProblems/369.lean is absent.

An unpinned current FormalConjectures file does exist at commit 6fbb54f24ccc2e64dcfaffc28c58950e377110d2. It formalizes the [n/2, n] variant with prime-factor bounds against \(n^ε\), but its theorem body is by sorry. The file cites an external plby/lean-proofs source at commit 1d7b3f00780b85ed0462e79a1cd5655, blob adb3f4a062a46766f615ac8e750aa46e846e38be. That source is evidence of a claimed formal proof artifact, not a replay performed here. Consequently, “formalized,” “formal proof source exists,” “proof compiles,” and “GCL mathematically certified” remain distinct statements. V2 ledger disposition, not this article, decides accepted coverage.

What remains

The most useful bounded successor is a source-reconciliation pass. Its input should be the frozen YAML entry and blob, the live page revision, the Yang TeX hash, the Balog–Wooley PDF, the exact FC-GDM revision, the current FC file hash, and the cited plby file hash. Its target should be a table matching every quantifier, interval endpoint, block indexing convention, smoothness bound, and formal status label across those sources.

Acceptance should require an explicit statement-equivalence or statement-variant verdict, a dated provenance chain, and a record that the FC-GDM snapshot lacks the file. Falsification is any mismatch in the universal quantifiers, the location interval, the base of the smoothness exponent, or the meaning of “sufficiently large.” The task should stop after that comparison and source capture; it should not launch an exhaustive Lean build or promote the catalogue label. Any mathematical certification, publication decision, or acceptance of the external proof belongs to the authorised Programme route.

Sources and claim boundaries

Primary links are the original issue, the frozen catalogue file, the live problem page, the Yang manuscript, and the Balog–Wooley paper. The cited current FC file is here, and the external formal source is here.

The original issue’s requested RESULT/1 citation is authenticated original RESULT/1. No future comment URL is inferred here. This article reports source evidence, a checked finite example, and unresolved status boundaries; it does not claim independent mathematical review, Lean replay, certification, publication, priority, or award status.

Catalogue and evidence provenance

  • Pinned catalogue source: teorth/erdosproblems at 536408716724.
  • GCL issue: #1639 — protected issue/assignment identity; its labels and results are live, not reproduced automatically here.
  • GCL triage: OUTSIDE_STRICT_OPEN_TRIAGE.
  • Formal-statement coverage: No statement was confirmed in GCL's pinned formal-provider snapshot. Presence does not mean the statement or proof has been checked.
  • External mathematical status: consult current problem page and its dated bibliography.

Back to catalogue · Compendium introduction