← All problems

Mondrian art problem defect-growth conjecture

Let d(n)d(n) denote the minimum defect among all Mondrian partitions of an n×nn\times n square. The conjecture is that lim⁡n→∞d(n)=∞\lim_{n\to\infty} d(n)=\infty; equivalently, the minimum defect must eventually exceed every fixed bound as the side length nn grows.

References

Source: arXiv

Progress summary

Refreshed
Open · progress reported

A 2026 preprint claims the proposed growth law is false by constructing infinitely many large squares with bounded defect, while related questions remain open.

The conjecture predicts that the minimum defect of Mondrian partitions grows with the side length. On October 13, 2026, Cooper O’Kuhn and Freeman Slaughter claimed infinite counterexample families, but the construction has not been independently verified.

Known results

  • Dalfó, Fiol, and López (2020) studied d(n)d(n), solved several small cases, and constructed families with d(n)/n2→0d(n)/n^2\to0.
  • Garrison, Seiler, and Semchankau (2026) claimed the general upper bound d(n)=O(n5/6)d(n)=O(n^{5/6}), improving the previously conjectured order n/log⁡nn/\log n.
  • Earlier work established restrictions and lower bounds for perfect partitions with d(n)=0d(n)=0.

October 13, 2026 bounded-defect construction

O’Kuhn and Slaughter give a family with defect at most 280280, beginning with n0=315n_0=315, n1=106785n_1=106785, and recurrence nj+2=340nj+1−njn_{j+2}=340n_{j+1}-n_j. They also claim a defect-560560 family meeting every divisibility class. This would refute the proposed growth behavior, but the preprint is unverified; the liminf-280280 conjecture and rectangular analogue remain open.

Current status (as of October 2026): The growth conjecture is claimed false by explicit bounded-defect families, but that claim lacks independent verification; the liminf-280280 and rectangular questions remain open.

Sources

Solutions 0

No solutions have been posted yet.