← All problems

Sierksma’s conjecture

For every pair of integers d≥1d\ge 1 and r≥2r\ge 2, every affine map a ⁣:Δ(d+1)(r−1)→Rda\colon \Delta_{(d+1)(r-1)}\to\mathbb{R}^d has at least ((r−1)!)d((r-1)!)^d unordered partitions of the vertex set of Δ(d+1)(r−1)\Delta_{(d+1)(r-1)} into rr nonempty parts V1,…,VrV_1,\ldots,V_r such that ⋂i=1rconv⁡(a(Vi))≠∅\bigcap_{i=1}^r\operatorname{conv}(a(V_i))\ne\varnothing.

Equivalent formulations
  1. Every set of (r−1)(d+1)+1(r-1)(d+1)+1 points in Rd\mathbb{R}^d has at least ((r−1)!)d((r-1)!)^d unordered partitions into rr parts whose convex hulls have a common point.

    source: Background summary supplied with the source

References

Source: arXiv

Progress summary

Refreshed
Open · progress reported

An unrefereed preprint reports a weaker lower bound for the number of partitions, but the conjectured bound remains unproved.

Sierksma’s conjecture says that every set of (r−1)(d+1)+1(r-1)(d+1)+1 points in Rd\mathbb{R}^d has at least ((r−1)!)d((r-1)!)^d unordered Tverberg partitions. It remains open in general.

Known results

  • Hell, 2008: proved the case d=2d=2, r=3r=3, giving at least 44 partitions.
  • Vučić–Živaljević and Hell: established weaker lower bounds when rr is a prime power.
  • The case d=1d=1 is settled for arbitrary rr.
  • The conjectured bound is sharp for known extremal configurations, but its universal validity is unproved.

October 2026 lower-bound preprint

Blagojević and Blagojević claim an explicit lower bound involving a prime-power component of rr, extending quantitative results beyond the prime-power regime. The bound is weaker than ((r−1)!)d((r-1)!)^d in general, so this is claimed progress rather than a resolution; the preprint is unrefereed.

Current status (as of October 2026): The conjecture is settled for d=1d=1 and for d=2d=2, r=3r=3; the general bound remains open, with a weaker non-prime-power lower bound claimed in an unrefereed preprint.

Sources

Solutions 0

No solutions have been posted yet.