Skip to content

Erdős Problem #17

Editorial depth: Reviewed explanatory article — evidence qualified.
Snapshot classification: open (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.

The question in plain language

A prime \(p\) is called a cluster prime when every even integer \(d\) from zero through \(p-3\) can be expressed as the difference of two primes not exceeding \(p\). Erdős Problem #17 asks whether infinitely many cluster primes exist. For each admissible even \(d\), the two witness primes may depend on \(d\); their upper bound \(p\) is essential.

A small example

A number \(d\) has the required representation exactly when some prime \(q\le p-d\) also has \(q+d\) prime. A finite sieve thus checks a candidate \(p\) without considering unrestricted differences.

The first noncluster prime is reported to be \(p=97\). The even difference \(d=88\) has no representation: any smaller prime \(q\) would have to satisfy \(q\le9\), leaving \(q=2,3,5,7\), but \(q+88=90,91,93,95\) are all composite. In contrast, the audit's exact sieve found every prime smaller than 97 to meet the cluster test. This is a bounded failure at one prime, not a refutation of infinitely many successful primes.

What is known

The maintained problem registry reports the existence question as OPEN and cites Erdős's 1995 problem statement, with later counting work by Benkoski, Erdős and Selfridge and by Elsholtz. The GCL source packet does not claim to have inspected those papers' complete arguments, so their reported bounds should not be promoted to independently verified theorems here.

GCL's current state

The exact source audit of the pinned Formal Conjectures file reconciles the even-difference condition and integer casting, proves the finite equivalence with the single-variable prime search, and reports an exhaustive sieve for primes through 101 with the missing-difference witness at 97. The formal provider's open statement has an unfilled proof, and the finite audit does not imply GCL mathematical certification.

What remains

The open question requires either infinitely many cluster primes or a proof that only finitely many exist. Neither finding one exceptional prime nor verifying many small examples constrains all sufficiently large primes. Any proposed sieve theorem must respect every even difference up to \(p-3\) for the same \(p\).

Sources and claim boundaries

Catalogue and evidence provenance

  • Pinned catalogue source: teorth/erdosproblems at 536408716724.
  • GCL issue: #1286 — 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.

Back to catalogue · Compendium introduction