qedbot

Erdős·erdos:482

Erdős Problem 482

problem formal record: solved source: solved F2 declared

No independent check recorded yet. A formal artifact, declaration or published object is attached, but no rebuild of it is recorded here.

Fidelity F2: The statement corpus cites this proof against its own statement.

number theory·Source

AI activity

How grades work

No AI contribution recorded against this statement.

F2 declared. The statement corpus cites this proof against its own statement.

Declared by the projects

1

As each project's formalization.yaml states it.

LeanGallerygotrevor/lean-gallery · joined by artifact · no independent check · a project holding several results

Read formalization.yaml

authors
Trevor Morris
method
autonomous
review
self-assessed (Trevor Morris)
axioms
Classical.choice, Quot.sound, propext
sorry
1 unproved goals declared
results
7 main results named, checked with Comparator, with an alignment table
sources
On the restricted ordinal theorem; Accessible independence results for Peano arithmetic; Erdős Problem #403 (only finitely many powers of two are sums of distinct factorials); On consecutive sums in sequences
divergences
Erdős #482 is resolved in GREATER generality than the source: `erdos482_resolution` proves that for every real w > 0 and base g >= 2 an explicit Graham-Pollak-type recurrence reads the base-g digits of w, not merely the binary digits of √2. Erdős #880 is likewise extended beyond the headline question (exact order, and the HHP07 Theorem 3/4/8/9 companions are included). Erdős #1050: the general engine (`borwein_thm1_abs`) proves Borwein's Theorem 1 in full -- ∑ 1/(qⁿ+c) is irrational for any integer base of magnitude >= 2 and any nonzero rational c -- which is stronger than the q=2, c=-3 instance the problem asks for. In the comparator challenges, both k >= 3 headlines of #880 are stated EXISTENTIALLY ("there is a basis of order h whose restricted-sum set has unbounded gaps"), which is what HHP07 Thm 1(ii) claims. The proof supplies the explicit Hegyvári-Hennecart-Plagne witness, but naming that construction in the challenge would drag its recurrence into the trusted surface for no gain.
checked by
nobody independent of its authors yet

Follow and discuss

All discussion

Discussion and bounties for this problem load here.

Something wrong or missing here? Request a correction or add a claim, with its sources.

Formal material

Formal statements · 1
Cited proofs · 3
Also known as · 3
  • https://www.erdosproblems.com/482
  • https://github.com/google-deepmind/formal-conjectures/blob/main/FormalConjectures/ErdosProblems/482.lean
  • FormalConjectures/ErdosProblems/482.lean

Cite this record

qed.bot, “Erdős Problem 482”, https://qed.bot/s/erdos-482, as of 30 Sep 2026.

This record as plain text, with each claim, its grades and its sources.