DEBRIEF · Claims Board
by Nervous Machine · generated 2026-11-14 from the mission ledgers

Claims by Classification

areaRegisteredChecks rerunReading guide writtenRead independentlyClaim rule met
Analysis: polynomial interpolationErdős #671····
Analysis: polynomialsErdős #1133····
Graph theory: Ramsey theoryErdős #550····
Number theory: arithmetic functionsErdős #1061····
Number theory: binomial coefficientsErdős #731····
Number theory: extremal problemsErdős #856····
Number theory: gaps between primes·Prime gaps ≤ 240 (Zhang)Prime gaps ≤ 186 (OpenAI)··
Number theory: multiplicative structure··Erdős #768··
Number theory: unit fractions··Erdős #306··
Set theory and hypergraphs: chromatic numberErdős #1177Erdős #593····
maintainer onlyindependent contributors

Individual claims and contributions

Each claim has its own repo. Status is defined in the main protocol.

claimstage · checkedwhat it needs nextcontributors
Erdős #306
The claimed solution of Erdős Problem #306 (sums of unit fractions with semiprime denominators)
Rests on an arXiv paper and a Lean development, built on an earlier conditional Lean development by another author.
Reading guide written
2 certified, 0 accepted · 2 open
Do the formal statements say what the problem asks?
a machine comparison of the claimant's main theorem with the third-party formal statement in google-deepmind/formal-conjectures, then a human read of any difference
Nervous Machine
last cycle 14 Nov 2026
Erdős #768
The claimed solution of Erdős Problem #768 (density of integers with a witness divisor for every prime factor)
Rests on an arXiv paper and a Lean development that vendors two other Lean projects.
Reading guide written
1 certified, 0 accepted · 2 open
Do the formal statements say what the problem asks?
two independent mathematicians read the formal statement against the problem as posed and sign, or refuse
Nervous Machine
last cycle 14 Nov 2026
Prime gaps ≤ 186 (OpenAI)
The claim that primes come within 186 of each other infinitely often
Rests on a Lean formalization with three stated assumptions, a numerical certificate (Python/FLINT) and a paper.
Reading guide written
5 certified, 5 accepted · 3 open
Do the formal statements say what a mathematician means?
a number theorist reads the definitional trace against the formal file and the paper and signs its checklist
Fredrik Prüzelius
last cycle 14 Nov 2026
Prime gaps ≤ 240 (Zhang)
The claim that primes come within 240 of each other infinitely often (Zhang)
Rests on an exact arithmetic certificate (Python) and a paper.
Checks rerun
3 certified, 1 accepted · 3 open
The analytic argument has no formal deduction and no referee report
a referee report or an independent expert read of the analytic sections
Nervous Machine
last cycle 8 Nov 2026
Erdős #1133
The claimed solution of Erdős Problem #1133 (polynomials agreeing with bounded data at most nodes)
Rests on a PDF write-up in a public repository; no formal deduction.
Registered
0 certified, 0 accepted · 2 open
No formal deduction exists for the claimed proof
a Lean formalization of the main theorem, by the claimant or anyone else, with the statement written by a third party
no contributions
last cycle 29 Apr 2026
Erdős #1061
The claimed solution of Erdős Problem #1061 (solutions of sigma(a) + sigma(b) = sigma(a+b))
Rests on an arXiv paper only; no formal deduction.
Registered
0 certified, 0 accepted · 2 open
No formal deduction exists for the claimed proof
a Lean formalization of the main theorem, by the claimant or anyone else, with the statement written by a third party
no contributions
last cycle 24 Jun 2026
Erdős #856
The claimed solution of Erdős Problem #856 (sets with no k elements sharing a pairwise least common multiple)
Rests on an Overleaf read-only document; no fixed file and no formal deduction.
Registered
0 certified, 0 accepted · 3 open
No formal deduction exists for the claimed proof
a Lean formalization of the main theorem, by the claimant or anyone else, with the statement written by a third party
no contributions
last cycle 15 Apr 2026
Erdős #550
The claimed solution of Erdős Problem #550 (Ramsey numbers of trees versus complete multipartite graphs)
Rests on an arXiv paper and a Lean development.
Registered
0 certified, 0 accepted · 3 open
The Lean development has not been built or kernel-checked by anyone outside the claimant
a pre-registered rerun: fetch at the pinned commit, build, confirm no sorry or admit, print the axioms
no contributions
last cycle 22 Jun 2026
Erdős #1177
The claimed solution of Erdős Problem #1177 (three statements about triple systems avoiding a fixed finite one)
Rests on an arXiv paper and a Lean development shared with the claim on problem 593.
Registered
0 certified, 0 accepted · 3 open
The Lean development has not been built or kernel-checked by anyone outside the claimant
a pre-registered rerun: fetch at the pinned commit, build, confirm no sorry or admit, print the axioms
no contributions
last cycle 23 Jun 2026
Erdős #671
The claimed solution of Erdős Problem #671 (Lagrange interpolation and the Lebesgue function)
Rests on two written arguments and two Lean files, one pinned by hash and one a live-editor link.
Registered
0 certified, 0 accepted · 4 open
The Lean development has not been built or kernel-checked by anyone outside the claimant
a pre-registered rerun: fetch at the pinned commit, build, confirm no sorry or admit, print the axioms
no contributions
last cycle 22 Jun 2026
Erdős #593
The claimed solutions of Erdős Problem #593 (finite triple systems forced by uncountable chromatic number)
Rests on two independent arXiv or PDF write-ups and two Lean developments, one per claimant.
Registered
0 certified, 0 accepted · 4 open
The Lean development has not been built or kernel-checked by anyone outside the claimant
a pre-registered rerun: fetch at the pinned commit, build, confirm no sorry or admit, print the axioms
no contributions
last cycle 23 Jun 2026
Erdős #731
The claimed solution of Erdős Problem #731 (the least non-divisor of the central binomial coefficient)
Rests on an arXiv paper and a Lean development, under the claimant's own formalisation of the word 'reasonable'.
Registered
0 certified, 0 accepted · 4 open
The Lean development has not been built or kernel-checked by anyone outside the claimant
a pre-registered rerun: fetch at the pinned commit, build, confirm no sorry or admit, print the axioms
no contributions
last cycle 27 Jun 2026