Skip to content

Documentary state: DRAFT_UNACCEPTED. This draft has NOT been editorially admitted. Mathematical claims require independent checking.

OAI-26 family 084: the geometric case of the Erdős similarity conjecture

Editorial status: DRAFT_UNACCEPTED. NO_CERTIFICATION. This article reports two source-pinned manuscript claims without independently verifying their proofs.

The question in plain language

The Erdős similarity conjecture concerns whether a set of real numbers with positive measure must contain a shifted and rescaled copy of a prescribed infinite pattern. For patterns whose points approach zero geometrically, the pinned OAI-26 overview claims that one can build sets occupying almost all of the unit interval while avoiding every nontrivial similar copy of the pattern.

The result is about an infinite geometric sequence. It does not establish the full Erdős similarity conjecture for arbitrary infinite patterns.

A small example

Fix a ratio \(q\) strictly between zero and one and consider

\[ E_q=\{q,q^2,q^3,\ldots\}. \]

A similar copy is obtained by replacing each point \(x\) with \(a+bx\), where \(a\) is a translation and \(b\ne0\) is a dilation. Negative values of \(b\) reflect the sequence; positive ones preserve its orientation.

The claimed construction gives a compact set \(K\) inside \([0,1]\) with Lebesgue measure as close to one as desired, but with no set \(a+bE_q\) contained in \(K\) for any \(a\) and nonzero \(b\). The absence of a copy is stronger than simply avoiding one particular translation or one particular scaling.

What is known

The protected Family 084 catalogue identifies two manuscripts: The geometric case of the Erdős similarity conjecture (October 5, 2026) and The dyadic case of the Erdős similarity conjecture (September 25, 2026).

The pinned publisher overview claims that for every fixed ratio \(q\in(0,1)\) and every desired measure deficit, there is a compact subset of the unit interval whose measure is correspondingly near one and which contains no translated, nontrivially dilated copy of the infinite sequence \(\{q^n:n\ge1\}\).

The statement covers dilations of either sign. The dyadic case, associated with a specific ratio, is related but does not alone imply the theorem for every ratio; the relationship between the two manuscripts requires a source audit.

The result supplies arbitrarily large-measure avoiders, not necessarily a single measure-one avoiding set. It is also not a theorem about avoiding all geometric ratios simultaneously with the same set.

GCL verification state

Source lock: openai/math@1581d4cbdd69a0faf34b2089614fe2a814316a6a. The protected catalogue identifies both source manuscripts; the pinned overview provides the claimed geometric-progression theorem.

This editorial pass has not independently reconstructed the measure-theoretic construction or checked avoidance across all translations and signed nonzero dilations. No Family 084-specific Lean proof, Comparator result, semantic source audit or MATHCERT certificate is established.

What remains

Read both pinned PDFs and identify their main theorems, exact quantifiers, and dependence of the avoiding set on \(q\) and the measure deficit. Verify compactness, measure estimates, and the exclusion of all translated and nontrivially dilated infinite patterns, including reflected copies.

Check how the dyadic construction extends to arbitrary ratios, whether the arguments use geometric decay quantitatively, and whether any exceptional ratios or hidden regularity assumptions occur in the proofs. Validate Markdown and rendered-compendium pages, record source-specific editorial criticism, and complete protected merge/readback.

Sources and claim boundaries

Boundary: NO_CERTIFICATION. The claimed geometric-progression result is publisher-reported, not independently certified by this documentary pass.


Indexed primary manuscript paths: