Erdős Problem #786
Editorial depth: Reviewed explanatory article — evidence qualified.
Snapshot classification: open (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
A multiplicative analogue of independent subset sums requires that equal products contain the same number of factors. Whether the factors can repeat is a critical distinction that changes the answer.
A subset A of positive integers is required to have the property that equality between products of r and s selected factors forces r=s. Does A have density arbitrarily close to one, or can A subset [1,N] have (1-o(1))N members? The exact definition MUST specify whether factors may repeat. Distinct-within-each-product and repetition-allowed versions are mathematically inequivalent.
A small example
Every number congruent to 2 modulo 4 has 2-adic valuation one. Therefore a product of r such numbers cannot equal a product of s such numbers unless r=s. The family has natural density one quarter.
Exact example A={2,6,10,...} numbers congruent 2 mod 4: v_2(a)=1 for every a, hence equality of products implies same number of factors even ALLOWING repetitions. Density 1/4. Conversely, A={2,4} satisfies 2*2=4 with factor counts 2 vs 1 when repetition is allowed, but this collision may vanish when every factor in each product must be distinct. It is an exact witness that conventions differ.
What is known
Maintained https://www.erdosproblems.com/786 OPEN, 11 April 2026 revision, expressly states repetitions-allowed case has negative density bounds from Erdős–Ruzsa–Sárközy [ERS73], with density <=1/2, whereas no-repetition version remains OPEN. Historical Erdős [Er80] says distinct factors and nevertheless attributes a negative result to Ruzsa; this is an unresolved source-semantic conflict, not a license to declare the intended problem solved. Original full [ERS73] proof and [Er80] pages not independently read. Terence Tao's public comments distinguish the linear-algebra reduction under unlimited repetition.
GCL's current state
The original authenticated RESULT/1 source return has historical disposition STATUS_CONFLICT. This status and the original research exceptions are preserved, not silently converted into a theorem verdict. The maintained Erdős source and source-specific mathematical definitions were reviewed, and this article is editorially reviewed as a source-exception-annotated explanation under EC-BULK-005-R10, effective only after the exact-hash protected admission merges.
Source interpretation and unresolved mathematical qualification: Allowing repeated prime factors versus distinct factors changes multiplication uniqueness; the source conflates conventions and its general density question remains open.
The explanation is accepted for publication of established background, examples, and explicitly attributed current claims. It does not establish or formally certify the disputed original theorem, external model-assisted proof, historical interpretation, or any unreplayed Lean kernel file. Original source-fidelity and proof-review research holds remain.
What remains
Successor: implement brute-force short-product collision checks under each convention separately, source-lock [Er80] p.114 and [ERS73] Theorem 2, and compare pinned formal statements if present. HOLD both asserted refutation and claimed proof for distinct-term problem pending source semantics.
The deployed MathJax has already been browser-checked; the genuine remaining obligation is source-fidelity and independent mathematical review of the distinct-factor interpretation.
Sources and claim boundaries
- Maintained Erdős Problem #786 and authenticated source-return record.
- Pinned teorth/erdosproblems source commit
536408716724a2402bdfa7c1d764bf83d64a8c89atdata/problems.yaml. Primary historical references are identified by their bibliography keys in the original return; any not directly inspected are explicitly unverified. - Source review: PARTIAL. Independent mathematical critical review: PENDING. Protected merge: NONE. Public browser MathJax: NOT VERIFIED. GCL theorem certification: NONE.
Qualified editorial critical review: R10 source-exception dossier. The source-semantic exception is visible and its original research hold remains. Publication classification: REVIEWED EXPLANATORY ARTICLE WITH UNRESOLVED SOURCE/PROOF QUALIFICATIONS. Independent theorem proof or Lean certification: NOT CLAIMED.
Catalogue and evidence provenance
- Pinned catalogue source: teorth/erdosproblems at
536408716724. - GCL issue: #2057 — protected issue/assignment identity; its labels and results are live, not reproduced automatically here.
- GCL triage: PROTECTED_READY_BACKLOG.
- Formal-statement coverage: A provider statement was included in the pinned snapshot. Presence does not mean the statement or proof has been checked.
- External mathematical status: consult current problem page and its dated bibliography.