← All problems

Dolgachev's essential-dimension and Cremona-dimension inequality

14 years old · documented by

Let GG be a finite group over the complex numbers. Write ed⁡(G)\operatorname{ed}(G) for the essential dimension of GG, and Crdim⁡(G)\operatorname{Crdim}(G) for the least integer nn such that GG embeds into the Cremona group Cr⁡(n)\operatorname{Cr}(n).

Dolgachev's inequality.

ed⁡(G)⩾Crdim⁡(G).\operatorname{ed}(G)\geqslant\operatorname{Crdim}(G).

The source attributes this claim to I. Dolgachev as unpublished. It compares the minimal dimension of a versal faithful GG-variety with the minimal dimension of affine space on which GG acts birationally; the source gives no resolution status.

References
  1. Alexander Duncan and Zinovy Reichstein, “Versality of algebraic group actions and rational points on twisted varieties”, arXiv:1109.6093 (2013).

Progress summary

Refreshed
Open

The inequality has supporting examples but no general proof or counterexample has been found.

Dolgachev proposed in a 2010 talk that, for every finite group over the complex numbers, essential dimension is at least Cremona dimension. The conjecture remains unresolved.

Known results

  • The inequality is verified for some finite pp-groups, groups generated by pseudo-reflections, and some simple or almost simple groups.
  • For elementary abelian groups, crd⁡k(2r)=rcd⁡k(2r)<ed⁡k(2r)=r\operatorname{crd}_{k}(2^{r})=\operatorname{rcd}_{k}(2^{r})<\operatorname{ed}_{k}(2^{r})=r.
  • ed⁡(A6)=3\operatorname{ed}(A_{6})=3, while ed⁡(A7)=ed⁡(S7)=4\operatorname{ed}(A_{7})=\operatorname{ed}(S_{7})=4.
  • 3≤ed⁡(PSL⁡2(F11))≤43\leq\operatorname{ed}(\operatorname{PSL}_{2}(\mathbb{F}_{11}))\leq4 remains unresolved.

August 2025 survey and candidate counterexamples

In a survey revised on August 4, 2025, Dolgachev stated that he did not know a proof and suggested seeking a counterexample. It identifies Sp⁡4(3)\operatorname{Sp}_{4}(3) as problematic if ed⁡C(Sp⁡4(3))=3\operatorname{ed}_{\mathbb{C}}(\operatorname{Sp}_{4}(3))=3, and discusses extraspecial groups p1+2np^{1+2n}; neither yields a reported counterexample.

Current status (as of October 2026): Supporting special cases are known, but the general inequality remains open, with no reported proof or counterexample.

Sources

Solutions 0

No solutions have been posted yet.