Search arXiv⌕ Search

arXiv · 2610.28124

Balanced subgraph counts and quasirandomness: a proof of the Janson--Sós conjecture

Abstract

Let $F$ be a finite simple graph with $m$ vertices and at least two edges, and let $p\in(0,1]$. We prove that a graph sequence is $p$-quasirandom if and only if the symmetrized counts of copies of $F$ across every balanced partition into $m$ classes have the asymptotics predicted by the random graph with edge probability $p$. This settles a conjecture of Janson and Sós for all such $F$, including graphs with isolated vertices. The proof completes their graph-limit reduction by a concavity argument for a sum of fractional powers determined by the degree sequence of $F$. We also show that a symmetrized subgraph kernel of positive mean can be additive in its coordinates only when the underlying graphon is constant almost everywhere. This rigidity result yields a quasirandomness criterion for generalized restricted subgraph counts, resolving a further conjecture of Janson.

Explore related subjects

Keep this discovery

Explore connections, maps & timelines

BibTeXRIS

Guoqing Wang, Linlin Wang, Jun Yue. 2026-10-21. Balanced subgraph counts and quasirandomness: a proof of the Janson--Sós conjecture. https://arxiv.org/abs/2610.28124

Cite the original work for its findings. Save a collection to share your selection of sources.

KEEP EXPLORING

Related papers

$2$-large sets are sets of nil-Bohr recurrence

A set $S$ of positive integers is $2$-large if every two-coloring contains arbitrarily long monochromatic arithmetic progressions with common difference in $S$. We prove that every such set meets every nil-Bohr neighborhood of $0$. This proves a conjecture of the first author and gives a sharp answer to a question of Host, Kra, and Maass. We also show that, for every fixed nil-Bohr neighborhood $B$ of $0$ and every $\varepsilon>0$, restricting the common difference to $B$ increases the extremal number for $k$-term arithmetic progressions by at most a factor of $1+\varepsilon$, provided $k$ is sufficiently large.

math.CO↗

On the Number of Iterations in a Kruskal-Type Process

Motivated by the Kruskal counting procedure, this paper investigates a combinatorial elimination process defined on a deck consisting of $m$ copies of each value $1,2,\ldots,n$. Starting from an arbitrary arrangement of the $mn$ cards, the process repeatedly removes a number of consecutive cards equal to the value of the first remaining card. This generates a sequence of elimination steps whose length determines $T$, representing the total number of iterations required to exhaust the deck. The main objective of this paper is to analyze an upper bound $σ(m,n)$ for $T$, valid for every initial configuration of the deck, and to show that this bound is attained by a specific configuration.

math.CO↗

Every tournament of dichromatic number $10 k^2$ has a vertex whose out-neighborhood has dichromatic number more than $k$

The dichromatic number of a tournament is the minimum number of colors needed to color the vertices so that no directed cycle is monochromatic. Nguyen, Scott, and Seymour (2025) asked whether every tournament with dichromatic number $2k$ has a vertex whose out-neighborhood induces a subtournament of dichromatic number at least $k$. We answer their question negatively by finding an example with dichromatic number $Ω(k^2)$. Furthermore, this quadratic bound is tight up to a constant factor; we show that for every integer $k \geq 1$, every tournament of dichromatic number $Ω(k^2)$ contains a vertex whose out-neighborhood induces a subtournament of dichromatic number more than $k$. This improves the previous best bound $\mathcal{O}(k^2\log k)$ due to Colinot and Newman (2026). To show this, we prove that every tournament of fractional dichromatic number $c$ has a dominating set of size at most $5c-1$. The proof uses Brouwer's fixed point theorem.

math.CO↗