Erdős Problem #876
No proof or disproof is known, and it isn't known whether a finite computation could settle it.
What the community database records, not our verdict. erdosproblems.com no longer shows statuses (why).
`A` is *sum-free* in the sense of Erdős: its elements are positive and no element is a sum of distinct smaller elements of `A`. The condition `0 < a` fits \(A\subset\mathbb{N}\) with \(\mathbb{N}=\{1,2,\ldots\}\) and makes the empty sum \(0\) harmless. A single smaller element cannot equal `a`, so only sums of \(r\geq 2\) elements matter. This differs from the usual `IsSumFree`: sums of any number of distinct smaller elements are forbidden, but \(a=b+b\) is allowed.
For an infinite set `A = {a₁ < a₂ < ⋯}`, `Nat.nth (· ∈ A) k` is \(a_{k+1}\), because the enumeration starts at index `0`. Hence `gap A n` is \(a_{n+1}-a_n\) for \(n\geq 1\). The value at `n = 0` is junk and is never used, since the statements below only concern large `n`.
`maxRecipSum` is the supremum of \(\sum_{a\in A}1/a\) over infinite sum-free sets `A`. Only convergent sums are taken (`HasSum`), so no junk value occurs. The set of values is nonempty (the powers of \(2\) give \(2\)) and bounded above by \(103\) [Er62c], so `sSup` is the true supremum, which may or may not be attained.
Attempts
submit one →Read our working report →
The live page is marked OPEN but lists 1 partial claimed proof (Liam Price using GPT-5.6 Sol Pro, submitted 2026-07-18) asserting that every infinite sum-free sequence satisfies \(\limsup_{n\to\infty}
SKIPPED