Threshold for minimal complements in finite groups

Question 10 · arXiv:2006.00534

arXiv Question high confidence— first stated 2020-12-30

Authors: Noga Alon, Noah Kravitz, Matt Larson

Status open high confidence

Question 10 asks whether the probability that a Bernoulli-p random subset of a finite group G is a minimal complement (or maximal supplement) exhibits a (sharp) threshold as a function of p = p(|G|). The source paper itself provides evidence for threshold behavior: Theorems 1 and 3 there show that for small p the random set is a minimal complement with high probability, and for large p it is not, but the existence and sharpness of a threshold are left open. No follow-up paper resolving this question was found in a broad search of the literature through May 2026.

Reviewer notes. No follow-up paper addressing the threshold question was found. The property of being a minimal complement is non-monotone, which makes standard sharp-threshold machinery (e.g., Friedgut–Kalai) not directly applicable; this non-monotonicity is explicitly noted in the source paper as the main obstacle. Searches on arXiv and via Kravitz's publication list returned no paper on this specific question. Status open with high confidence.

Auto-reviewed 2026-05-15 with claude-sonnet-4-6 (web search enabled).

Question. How does the probability that a random subset of a finite group $G$ with parameter $p = p(|G|)$ is a minimal complement (or maximal supplement) vary with $p$? Is there a (sharp) threshold?

Context

Although being a minimal complement is not a monotone property, Theorems 1 and 3 together show that for small $p$ a random set is with high probability a minimal complement, and for large $p$ it is with high probability not a minimal complement, suggesting threshold behavior.

Source paper

Inverse problems for minimal complements and maximal supplements
Noga Alon, Noah Kravitz, Matt Larson · 2020-12-30
https://arxiv.org/abs/2006.00534 PDF source