Erdős Problem #1202
Editorial depth: Authored explanatory DRAFT — NOT editorially accepted.
Snapshot classification: solved (not automatically the live result).
Subject areas: number theory, primes.
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.
UNACCEPTED DRAFT — mathematical editorial review pending
This material is a provisional, source-qualified manuscript published for line-by-line scrutiny and real HTML/LaTeX testing. It has NOT received separate GCL editorial acceptance or independent mathematical certification. Its presence on this website is not evidence of a solution, reviewed coverage, or final publication.
Review the underlying draft on GitHub.
The question in plain language
Erdős Problem #1202 has maintained status EXTERNALLY DISPROVED — SOURCE FIDELITY PENDING at this audit.
Fix positive \(\epsilon,\eta\) and distinct primes \(p_i<n^{1-\epsilon}\), choosing for each a subset \(A_i\) of exactly \((p_i-1)/2\) residue classes. The displayed claim asks whether some fixed \(k\) makes the set of positive \(m\le n\) avoiding all the selected residues have size at most \(\epsilon n\) regardless of all choices of primes and residue classes. The introductory parameter \(\eta\) is not present in the displayed conclusion, a genuine unused parameter requiring original-source comparison. Choosing congruence classes at random or assuming they are evenly spread is not a legitimate replacement for the universal adversarial quantifier.
A small example
For \(n=60\) and primes \(3,5,7\), each forbidden family has \((p-1)/2\) residues (1,2,3 classes). Exhaustively check \(3\binom52\binom73=1050\) choices; the count of surviving positive \(m\le60\) ranges between 11 and 17. One maximizing witness removes residues \(\{0\}\bmod3\), \(\{0,1\}\bmod5\), and \(\{4,5,6\}\bmod7\), leaving 17 integers. This tiny three-prime example is not a counterexample to the existence of some larger \(k\) valid asymptotically.
What is known
The maintained https://www.erdosproblems.com/1202 now marks the claim SOLVED IN THE NEGATIVE, by Price and GPT-5.4 Pro (April 2026). The reported construction chooses each forbidden \(A_i\) as an interval of residues while preserving a long progression of integers that survives all restrictions. It reports arbitrarily many primes in a comparable scale \(m<p_i\le2m\) and survivor proportion at least \(1/2-c\), contradicting any uniform bound with sufficiently small \(\epsilon\). The known large sieve argument supplies a positive result only for a more restricted prime-size regime (reported \(p_k<n^{1/2}\)). Independent mathematical certification of the recent counterexample, including the exact quantifiers and comparison \(\epsilon/\eta\), remains PENDING.
GCL's current state
Authenticated original issue-bound research return. The bounded mathematical fixture was independently replayed through ci/erdos_compendium_007_phase24_checks.py. All external advanced/general theorem statements remain attributed to sources, not independently certified. This is a local UNACCEPTED editorial manuscript.
What remains
Review original Er80 p.107, Green Problem 44, and Price/GPT-5.4 Pro 2026 argument for the fixed \(\epsilon,\eta,k\) quantifier order, the unused eta parameter, and the survivor progression after adversarial residue choices. Independently reproduce the construction and source-check the large-sieve threshold; do not admit it merely from online solved status.
Independent source fidelity and critical mathematical review, protected publication, actual public deployment and V2 admission remain separate uncompleted obligations.
Sources and claim boundaries
https://www.erdosproblems.com/1202 ; [Er80,p.107], Green Problem 44; 2026 Price/GPT-5.4 Pro external proof.
Exact finite receipt: governance/erdos_compendium/batches/EC-BULK-007_FINITE_CHECKS_PHASE24.json. Original source issue frozen SHA retained; no independent certificate claimed.
Catalogue and evidence provenance
- Pinned catalogue source: teorth/erdosproblems at
536408716724. - GCL issue: #2473 — 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.