Ramos conjecture for two hyperplanes

Let Δ(j,2)\Delta(j,2) denote the least integer dd such that every collection of jj absolutely continuous probability measures on Rd\mathbb{R}^d can be simultaneously equipartitioned by two affine hyperplanes, meaning that each of the four regions determined by the hyperplanes has measure 1/41/4 for every measure. The Ramos conjecture asserts that Δ(j,2)=⌊3j+12⌋\Delta(j,2)=\left\lfloor\frac{3j+1}{2}\right\rfloor (in particular, for every integer n≥2n\ge 2, every collection of 4n−24n-2 such measures in R6n−3\mathbb{R}^{6n-3} admits a common two-hyperplane equipartition).

References

Primary source

arXiv

Additional references

Progress summary

Refreshed
Claimed solved

A September 2026 paper claims to disprove the conjecture in every dimension covered by the two-hyperplane Gaussian formulation.

The Ramos conjecture predicts a specific dimension threshold for partitioning measures equally with two affine hyperplanes. Florian Frick’s new paper claims counterexamples in every integer dimension parameter n≥2n\ge 2, so the conjecture is false if the construction is correct.

Known results

  • A 2015 critical review reported rigorously established threshold values including Δ(j,2)=12(3j+1)\Delta(j,2)=\frac{1}{2}(3j+1) when j−1j-1 is a power of 22 and j≥5j\ge 5, while identifying gaps in several earlier claimed proofs.

September 23, 2026 counterexample

On September 23, 2026, Frick’s article Counterexamples to the Ramos conjecture for two hyperplanes claimed that, for every n≥2n\ge 2, 4n−24n-2 smooth nondegenerate Gaussian measures in R6n−3\mathbb{R}^{6n-3} have no common equipartition, giving Δ(4n−2,2)≥6n−2\Delta(4n-2,2)\ge 6n-2. Combined with upper bounds, it claims Δ(2s−2,2)=3⋅2s−1−2\Delta(2^s-2,2)=3\cdot2^{s-1}-2 for s≥3s\ge 3, including Δ(6,2)=10\Delta(6,2)=10, and analogous perpendicular-hyperplane results.

Current status (as of September 2026): The conjecture is claimed false for the Gaussian two-hyperplane formulation, with several exact threshold values claimed, but the new paper’s conclusions remain unverified.

Sources

Solutions 0

No solutions have been posted yet.