Skip to content

Erdős Problem #206

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

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 \(x>0\), let \(R_n(x)\) be the largest sum below \(x\) of precisely \(n\) distinct positive unit fractions \(1/m_i\). Erdős and Graham asked whether, for almost every real \(x\), sufficiently large best sums are formed from the preceding best sum by adding the largest new unit fraction still keeping the result strictly below \(x\).

The words almost every mean outside a Lebesgue measure-zero exceptional set. The conjecture is weaker than claiming that the usual greedy algorithm is optimal at every stage, because finitely many initial failures are allowed.

A small example

For \(x=11/24\), the best one-term underapproximation is \(R_1(x)=1/3\). The best two-term sum, however, is

\[ R_2(11/24)=\frac14+\frac15=\frac9{20}, \]

which is larger than extending \(1/3\) by its next admissible new fraction: \(1/3+1/9=4/9\). Both lie strictly below \(11/24\). This shows a finite-stage greedy failure, not that this particular rational is never eventually greedy.

Distinct denominators and the strict comparison with \(x\) are both necessary. Permitting repeated denominators or equality with \(x\) changes the optimization problem.

What is known

DISPROVED in the strongest measure-theoretic sense. Vjekoslav Kovač, On eventually greedy best underapproximations by Egyptian fractions, Journal of Number Theory 268 (2025), 39–48, proves that the set of real \(x>0\) whose best underapproximations are eventually greedy has Lebesgue measure zero.

The negative result does not provide a simple explicit irrational counterexample: the maintained Erdős #206 notes that finding an explicit irrational which is not eventually greedy remains open. For many rational numbers, including \(x=1/m\), eventual or complete greedy optimality is known; almost-all failure is consistent with all rationals occupying a measure-zero set.

The paper is also available as arXiv:2406.07218. Its full metric proof is source-attributed, not independently recreated.

GCL's current state

Pinned and maintained status: DISPROVED; original issue #1475. The pinned provider contains a #206 statement but an independently checked theorem-level mapping and Lean proof replay are still outstanding.

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 enumerates exact reduced rational unit-fraction sums of length one and two with denominators at most 30. Acceptance: verify the values \(R_1(11/24)=1/3\) and \(R_2(11/24)=9/20\) with explicit finite-exhaustiveness arguments, check the best new unit fraction after \(1/3\) is \(1/9\), and preserve strictness below \(x\). Do not infer an almost-everywhere theorem from rational enumeration.

Sources and claim boundaries

No new measure-theoretic proof, explicit irrational counterexample or editorial admission is claimed.

Catalogue and evidence provenance

  • Pinned catalogue source: teorth/erdosproblems at 536408716724.
  • GCL issue: #1475 — protected issue/assignment identity; its labels and results are live, not reproduced automatically here.
  • GCL triage: OUTSIDE_STRICT_OPEN_TRIAGE.
  • 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