Sharp Ambiguity Bounds for Surviving Conditional Odds Ratios
Manuscript 4 October 2026 · Online 4 October 2026
Abstract
Fix the support of a binary \(d\)-way probability table, its feasible interior one-way margins, and every conditional pairwise odds ratio whose four cells remain positive. The largest possible dimension of the resulting family is \(\lceil 2^{d+1}/3\rceil-d-1\) for \(d\ge2\), even with uniform margins. For \(d\ge3\), equality holds precisely on classical extremal square-free layer sets, up to cube automorphisms. These extremizers have no surviving ratios. We also give a four-variable uniform-margin family of dimension five with one genuinely surviving ratio. The proof combines classical mixed coordinates with a pivot-deletion lemma and the Kostochka–Johnson–Entringer vertex theorem. Two exact implementations verify all 65,805 nonempty supports through dimension four. This does not classify arbitrary collapsed marginal ratios or resolve the broad AIM specification question; priority is undetermined.
Record
- Affiliation
- Mercury Software GmbH
- Contact
- [email protected] · GitHub
- Result
- Sharp conditional-ratio theorem
- Categories
- math.ST · math.CO
- Manuscript
- 4 October 2026
- Online release
- 4 October 2026
- Version
- 1.0
- License
- Creative Commons Attribution 4.0 International
Files and verification
The PDF is the canonical reading copy. The source archive contains the LaTeX manuscript, bibliography, reproducibility material, and audit documents without build artefacts.
Citation
Alper Ferudun, “Sharp Ambiguity Bounds for Surviving Conditional Odds Ratios,” EulerSolve Research Papers, AIM-COMPUTATION-0020, 2026. https://doi.org/10.5281/zenodo.23130679.
BibTeX
@misc{Ferudun2026OddsRatioAmbiguity,
author = {Ferudun, Alper},
title = {Sharp Ambiguity Bounds for Surviving Conditional Odds Ratios},
year = {2026},
howpublished = {EulerSolve Research Papers},
url = {https://eulersolve.org/papers/aim-computation-0020/},
doi = {10.5281/zenodo.23130679},
note = {AIM-COMPUTATION-0020; unrefereed preprint}
}