Parity obstruction in the minimum-determinant problem for Latin squares
A Mathematics Stack Exchange question posted on 3 August 2014 asks when the standard divisibility lower bound for determinants of Latin square matrices is attained. For an Latin square with entries , let
For which positive integers does there exist such an with ? The question conjectures that are the only orders for which this minimum cannot be attained.
- Result
- Proved(see note)
- Status
- Partial result
- AI contribution
- AI-discovered
- Method
- Argument
- Field
- Latin squares; determinant divisibility
- Posed by
- Mathematics Stack Exchange user "Peter"
- Year posed
- 2014
- Years open
- 12y
- Solved
- 2026-04-22
- Model
- GPT-5.4
- Vendor
- OpenAI
- Collaborators
- —
- Verification
- Unreviewed
- Publication
- Announced
- Significance
- 10 / 100
- Disclosed cost
- —
- Wikipedia
- No dedicated article
What was actually shown
For even , let . The work proves that is even exactly when the stronger centered divisibility holds. For , this is equivalent to ; for , it is equivalent to . An explicit family gives odd for every , . This removes a universal extra-factor-two obstruction, but it does not prove . Exact minimum attainment and the separate singularity question remain open.
What the AI did
Under the author's direction, OpenAI's ChatGPT, using the GPT-5.4 model, generated the central mathematical development of this work, including the ordinary-to-centered determinant reduction, the exact binary rank and adjugate criteria governing the additional factor of two, and the all-order construction producing an odd determinant quotient for every , . It also assisted with the development of the exact verification code and the manuscript. The author selected the research direction, checked the mathematical derivations and certified outputs, established the public claim boundaries, commissioned adversarial reviews, and takes responsibility for the final content.
Verification
The paper and public repository contain complete proofs, exact certified datasets, and a deterministic verifier that currently passes all 12 public artifacts. The release also received an artifact-oriented adversarial audit. These checks establish internal consistency and reproducibility, not independent expert endorsement of the headline theorem; no domain expert has yet endorsed it. The appropriate VibeMathed verification label is therefore Unreviewed.
Sources
Submitted by VividMarten473 on