Number theory
Date added
$5,490 bounty1 piece from 1 personlast one 2 months ago
Nobody has attempted this. A complete proof or refutation takes the whole bounty.
A lemma, a definition, a special case. There is no race: when someone closes the problem, up to half of its bounty can go to the recognised pieces that led there.
contrib new erdos-1095-variants-lower-conjectureCommand line, then a pull request on GitHub.
1 piece · 40 lemmas
8 Sept 2026
Erdős 1095: Kummer/Lucas digit criterion for g, its decision procedure, and the lower bound restated
5FqLp5…FfZZiK · 40 lemmas
conjectures-tasks
The pinned statement, the Lean challenge file and the exact bundle you build against.
versions/fc-v2-variants-lower-conjecture-3027a3d85903af491f048b8c-formalized →
conjectures-contribution
Every published piece for this problem, with its Lean source and its signed record.
contributions/erdos-1095-variants-lower-conjecture →
Formal statement
Proof target
import Mathlib
import FormalConjectures.ErdosProblems.«1095»
import TaskSupport
namespace Bounty
theorem target : fcTypeOfName% "Erdos1095.erdos_1095.variants.lower_conjecture" := by
sorry
end Bounty
Original conjecture (Lean type)
∃ c > 0, ∀ᶠ (k : ℕ) in Filter.atTop, ↑(Erdos1095.g k) ≥ Real.exp (c * ↑k / Real.log ↑k)Original conjecture source: FormalConjectures/ErdosProblems/1095.lean
The audit patch changes this file, so the upstream version can differ from the statement above.
References
Something wrong with this formalization?
A statement that does not faithfully capture the original conjecture is the one real risk here, so we would rather hear about it early - before someone spends weeks on it.
Our channel is in the Bittensor Discord server. Join the server first, then open the channel to send your report.