Research notes on every problem, and a library of the papers behind them. Built from the open erdos repository.
Updated
Source. The accepted proof file Main.lean of Conjectures.io record
e93a2766-4c70-4564-b565-d0c556f35929, theorem target at line 14040,
reducing through erdos18b_of_weighted_dyadic_mixing (line 2580) to
weighted_dyadic_mixing (line 14011); provenance and acceptance on the
source record. The target and
its reduction were read; the body was not read; nothing was built here.
Statement
For a practical number let be the least number of distinct divisors of that always suffice: the maximum over of the least size of a set of divisors of summing to . For every there is such that
that is, .
Formal statement
True ↔ ∀ (ε : ℝ), 0 < ε → ∀ᶠ (n : ℕ) in Filter.atTop, ↑(Erdos18.practicalH n.factorial) < ↑n ^ ε,
with Erdos18.practicalH n the Finset.sup over Finset.Icc 1 n of the
sInf of the sizes of divisor sets D ⊆ n.divisors with m ∈ subsetSums D
(catalog file on main). The cast ↑(practicalH n.factorial) is to the
reals and ↑n ^ ε is the real power; ∀ᶠ … in Filter.atTop on ℕ is "for
all sufficiently large n"; the True ↔ prefix is the catalog's answer
marker filled affirmatively, so proving the iff proves the right side.
How the file closes the target (structure only)
exact_target_type (line 4) proves by rfl that fcTypeOfName% "Erdos18.erdos_18b" is the displayed proposition. WeightedDyadicMixing
(line 2472) is, as the declaration reads, a property of finite nonempty sets
of odd naturals: for every there are , and
such that whenever and the uniform measure of on
puts mass at most on every class for
, the -fold product cyclicProductPow of that measure on
has discrete Fourier transform of modulus at most
at every odd frequency (the auxiliary definitions cyclicUniformNatSet and
cyclicProductPow were not read). erdos18b_of_weighted_dyadic_mixing
(line 2580) derives the target from this property, weighted_dyadic_mixing
(line 14011) proves the property, and target (line 14040) combines the
two. The site's review states that the mixing premise is proved within the
file and that "conditional intermediate lemmas are not residual assumptions
of the final result."
Dependencies
The catalog's Erdos18.practicalH and Erdos18.factorial_isPractical
(proved in the catalog file) and Mathlib through the trusted challenge
environment; axiom closure within propext, Quot.sound and
Classical.choice by the site's audit.
Standing
Documented independent acceptance of the formal statement by the bounty site: kernel check, review, certification and payment. Not refereed; a single kernel; no fresh replay by the reviewers; not built here, so no local kernel credit. The site owner's post on the erdosproblems.com proof-claims tab of 27 September 2026 is explicitly not a verification.
Bears on
- Problem 18: proves the second question exactly and only it; the first and third questions are untouched.