Skip to content

Documentary state: EDITORIALLY_ADMITTED. The manuscript has a protected editorial admission, not a theorem certification.

OAI-26 family 025: Short Egyptian fractions

Editorial status: EDITORIALLY_ADMITTED. Mathematical status: NO_CERTIFICATION. This explanation records claims from the pinned source, not independently checked mathematical proofs.

The question in plain language

An Egyptian fraction writes a positive rational number as a sum of reciprocals of distinct positive integers. For instance, one half can be written as one third plus one sixth. How few terms suffice for any fraction with a prescribed denominator?

The September 25, 2026 manuscript Short Egyptian fractions claims to settle an Erdős conjecture: for each sufficiently large denominator \(b\), every numerator \(1\le a<b\) admits a distinct-unit-fraction expansion with at most a constant times \(\log\log b\) terms.

A small example

For the fraction \(5/6\), the identity

\[ \frac56=\frac12+\frac13 \]

uses two distinct positive unit fractions. The requirement that denominators differ prevents simply repeating the same reciprocal. The challenge is uniformity: one must find short representations simultaneously for every numerator with the same large denominator.

A fraction such as \(1/b\) already has a one-term expansion. The worst case, not the easiest numerator, governs the conjecture.

What is known

The pinned primary PDF is a 33-page manuscript whose displayed Theorem 1.1 defines \(N(a,b)\) as the minimum number of distinct unit fractions with denominators at least two representing \(a/b\), and \(N(b)\) as the maximum over \(1\le a<b\). It claims absolute constants \(c_1,c_2>0\) and \(b_0\) such that, for \(b\ge b_0\),

\[ c_1\log\log b\le N(b)\le c_2\log\log b. \]

The lower-order bound concerns the worst numerator. It is not a claim that every individual \(a/b\) requires this many terms. The source explicitly permits arbitrarily large denominators in the Egyptian-fraction expansion; an additional bound on their size is not part of Theorem 1.1.

Its introduction distinguishes earlier larger upper bounds from the new claimed double-logarithmic order. Its Corollary 1.2 also states that the number \(F(k)\) of expansions of one into exactly \(k\) distinct unit fractions satisfies \(\log\log F(k)\) of order \(k\) for sufficiently large \(k\). The abstract further claims a doubly exponential scale for the least denominator at least two excluded from all such expansions of length \(k\). These counting and prescribed-denominator results must not be conflated with the main uniform-length bound.

GCL verification state

The primary source was retrieved from openai/math@1581d4cbdd69a0faf34b2089614fe2a814316a6a: Git blob 380f422b0dccf4142e55d40045cbe254776d9c17, 530,973 bytes, 33 PDF pages. Its abstract and opening pages, including Theorem 1.1 and Corollary 1.2, were examined directly.

This establishes source-statement fidelity only. The proof's dense-numerator reduction, uniform divisor-moment bounds, residue-discrepancy construction and propagation of exceptional sets have not been independently verified. There is no demonstrated paper-to-Lean correspondence, Comparator replay, or MATHCERT certificate.

A bounded source-critical review and the exact-source-hash editorial admission were recorded through protected PR #3008. The admission establishes a reviewed exposition with explicit limits; it does not validate the manuscript's proof.

What remains

An independent mathematical critic should verify each step transferring a dense collection of short expansions to every numerator, the divisor-moment argument, construction of usable divisors with small residues and treatment of exceptional sets. Separately check the enumeration of expansions of one and the claimed prescribed-denominator consequence.

The admitted editorial scope covers attribution, quantifiers, a bounded source critic, and exact-manuscript-hash registration. Independent verification of the proof's dense-numerator reduction, divisor-moment estimates, discrepancy construction, exceptional-set propagation, and counting consequences remains outstanding. Editorial acceptance does not establish that the mathematics is correct.

Sources and claim boundaries

Boundary: NO_CERTIFICATION. This is an explanation of a pinned manuscript, not a verified proof of the Erdős conjecture.


Indexed primary manuscript paths:

  • Short Egyptian fractions — preprints/Short-Egyptian-fractions-September-25-2026/Short-Egyptian-fractions-September-25-2026.pdf