Erdős Problem #124
Editorial depth: Reviewed explanatory article — evidence qualified.
Snapshot classification: open (not automatically the live result).
Subject areas: number theory, base representations, complete sequences.
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
For integers \(d\ge3\) and \(k\ge0\), let \(P(d,k)\) consist of numbers formed by sums of distinct powers \(d^j\) with \(j\ge k\). Thus \(P(d,0)\) is the set of nonnegative integers whose digits in base \(d\) are only zero or one; \(P(d,k)\) contains multiples of \(d^k\) formed with the same digit restriction.
For a finite set of distinct bases \(3\le d_1<\cdots<d_r\), consider the reciprocal condition
The first question asks whether every sufficiently large integer belongs to the sumset \(P(d_1,0)+\cdots+P(d_r,0)\). The second, harder question adds \(\gcd(d_1,\ldots,d_r)=1\) and asks for the same conclusion with \(P(d_i,k)\) for every fixed \(k\ge1\).
A small example
The bases \(3,4,7\) satisfy
Burr, Erdős, Graham and Li proved the positive-shift \(k\ge1\) result for exactly these bases. This is a special case, not a solution for all admissible finite lists.
If the bases have common divisor \(g>1\), every element of \(P(d_i,k)\) for \(k\ge1\) is divisible by \(g\), so the sumset cannot represent all sufficiently large integers. This proves why the gcd hypothesis in the second question matters.
What is known
The first version with \(k=0\) has a reported affirmative Lean proof by an Aristotle agent coordinated by Boris Alexeev; the second, with \(k\ge1\) and coprime bases, remains OPEN in the maintained catalogue. The 1996 primary paper is Burr, Erdős, Graham and Li, Complete sequences of sets of integer powers, Acta Arithmetica 77, 133–138. The source's stronger example \((3,4,7)\) should not be confused with the full conjecture.
Source transcription defect: the live Erdős #124, edited December 2025, prints \(d_r\) in the denominator of every term of the introductory sum. Its own explanatory paragraph instead uses \(d_i\), and the protected Formal Conjectures 124 declaration independently records the indexed sum with \(d_i\). These are different mathematical conditions: for \((3,4,7)\) the printed repeated-last-base sum is \(3/(7-1)=1/2\), whereas the indexed sum is one. The indexed version is used here but the discrepancy must remain recorded.
The protected formal declaration contains separate branches for zero and nonzero digit cutoffs; its declarations carry proof holes. Reports of proofs or formalizations elsewhere are not equivalent to kernel verification under GCL's frozen source lock.
GCL's current state
The pinned catalogue marks this entry OPEN, correctly reflecting the unresolved second question. The original issue is #1393. The source-condition discrepancy is an explicit fidelity risk, not an excuse to replace the protected formula silently.
The original issue contains an authenticated RESULT/1 source report. This is unadjudicated source evidence only, not independent mathematical certification or editorial admission.
What remains
A bounded task is to formalize both reciprocal conditions as separate statements and prove they differ on \((3,4,7)\); then implement exact finite base-digit membership and sumset enumeration for \(P(3,k)+P(4,k)+P(7,k)\) with \(k=0,1\). Acceptance: no base-digit coefficient exceeds one, the cutoff \(k\) is enforced, and the gcd obstruction is checked for shared-divisor bases. Stop at a finite semantics and transcription audit, not an asymptotic representation theorem.
Sources and claim boundaries
- Burr–Erdős–Graham–Li, Acta Arithmetica 77 (1996), primary bibliographic and special-case source.
- Erdős #124, two questions and ambiguous printed summation.
- Pinned formal source; GCL issue #1393.
No all-cutoff theorem, independently checked Lean proof, or editorial admission is claimed.
Catalogue and evidence provenance
- Pinned catalogue source: teorth/erdosproblems at
536408716724. - GCL issue: #1393 — 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.