← All problems

Sharp thin-shell inequality for log-concave measures

22 years old · documented by

Suppose μ\mu is log-concave and isotropic. Is it true that

∫Rn(∣x∣−n)2 dμ(x)≤C\int_{\mathbb{R}^n} \left(|x|-\sqrt{n}\right)^2\,d\mu(x)\leq C

for a universal constant C>0C>0?

References

Progress summary

Refreshed
Open · solution reported

A newly available proof establishes the exact thin-shell variance constant as eight, with equality examples, so this problem is now settled.

The problem asks for the smallest universal constant in the variance bound for isotropic log-concave measures. The broader bounded-width thin-shell conjecture was resolved in 2025 by Klartag and Lehec, but the sharp variance constant is addressed separately by the result below.

Known results

  • Klartag and Lehec, 2025: proved bounded thin-shell width, namely a dimension-free bound for isotropic log-concave measures.

July 2026 sharp-constant proof

Theorem 1.1 proves, for every isotropic log-concave X∈RnX\in\mathbb{R}^n,

Var⁡(∣X∣2)≤8n.\operatorname{Var}(|X|^2)\leq 8n.

The constant 88 is optimal: equality holds for independent standard centered exponential coordinates. Among uniform measures on convex bodies, equality is attained by the regular simplex in isotropic position.

Current status (as of July 2026): The optimal universal constant is settled as 88, with explicit equality cases; no substantive part of the stated problem remains open.

Sources

Solutions 0

No solutions have been posted yet.