arXiv++ Combinatorics

For a $Δ$-regular connected graph $H$ we obtain the upper tail large deviations speed and rate function for the number of copies of $H$ in $\mathbb{G}_n^d$, a random $d$-regular graph on $n$ vertices, in the regime $np^{Δ/2} \gg (\log n)^{1/(v_H-2)}$ and $p \ll 1$, where $p:=d/n$ and $v_H$ is the number of vertices, and also identify the typical structure of $\mathbb{G}_n^d$ conditioned on this upper tail event. For $Δ\geq 3$, the speed and the rate function are same with those in the Erdős-Rényi setting. In contrast, for $Δ=2$ (i.e.~in case of cycles), due to the degree constraint in $\mathbb{G}_n^d$, although speed remains the same the rate function turns out to be different. Moreover, conditioned on the upper tail event of cycle counts in $\mathbb{G}_n^d$, we show that the random graph contains a disjoint collection of clique and biclique like structures with probability approaching to one, a stark contrast compared to the Erdős-Rényi setting. In the complementary regime, i.e.~for $1 \ll n^{v_H}p^{e_H} \ll (\log n)^{α^*_H/(α^*_H-1)}$, where $e_H$ is the number of edges in $H$, $α^*_H$ is the fractional independence number of $H$, and $H$ is a strictly balanced graph (and thus including the regular case), we show that the upper tail large deviations speed and rate function for the number of copies of $H$ match with those in the Erdős-Rényi setting. In this regime, we further show that almost all of the unlabeled copies of $H$ conditioned on the upper tail event are edge disjoint.

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.

We prove that every spanning subgraph $H$ of an $n$-vertex $(p,β)$-bijumbled graph, satisfying $δ(H)\geq(1/2+\varepsilon)pn$, where $\varepsilon > 0$ is arbitrarily small yet fixed, admits a rainbow Hamilton cycle under any globally $μpn$-bounded colouring, provided that $β\leq cpn$ and $pn\geq M$, where $μ$, $c$, and $M$ are positive constants depending solely on $\varepsilon$; in particular, $pn$ may be taken to be a sufficiently large constant. This answers a question of Coulson, Keevash, Perarnau and Yepremyan for random graphs and extends it to deterministic pseudorandom hosts. In fact, we prove that $H$ admits at least $(apn)^n$ rainbow Hamilton cycles, where $a > 0$ is a constant depending solely on $\varepsilon$. If $pn\geq D\log n$, then, upon fixing a coloured $H$, retaining each edge independently with probability $D\log n/(pn)$ preserves rainbow Hamiltonicity asymptotically almost surely. The logarithmic degree requirement is needed only for this percolation conclusion. All three conclusions also hold with rainbowness replaced by avoidance of prescribed pairs of edges, provided that each edge has at most $μpn$ conflicting partners. Our proofs use spread measures and the lopsided local lemma.

We study the graphon variational problem that governs lower tail large deviations of the number of copies of a fixed nonbipartite graph $H$ in a sparse Erdős--Rényi random graph. Our main new insight is that any non-constant minimizer of the variational problem exhibits a discontinuous jump away from the constant graphon. We deduce that the replica-symmetric and symmetry-breaking regimes are separated by a single critical parameter and that the phase transition is first order. In the supercritical regime $p \gg n^{-1/m_2(H)}$, we transfer these results to the conditioned random graph, showing that the edge density is a discontinuous order parameter and that the conditioned graph concentrates in cut distance near the full minimizer set.

We determine the exponential asymptotics of the number of connected labelled graphs with finitely many vertex types, prescribed numbers of vertices of each type, and prescribed edge counts between each pair of types. Under an irreducibility and strict surplus condition, we express the count for linear surplus of edges as the unrestricted count multiplied by an explicit exponential connectivity factor, up to a multiplicative error \(\exp\{o(n)\}\). The factor is characterized by a finite-dimensional nonlinear system whose solution exists and is unique. We use a probabilistic method, exploiting large-deviation principles and asymptotic estimates for connectedness probabilities to relate the counting problem to the emergence of giant components in suitably tuned supercritical random graphs. For connected bipartite graphs, we determine the exponential rate throughout the regime of positive linear excess and fixed positive limiting proportions of the two vertex classes. In the parameter ranges where the existing upper and lower bounds apply, this identifies the exponential rate left undetermined by those estimates.

Let $H_d(n,M)$ be a $d$-uniform hypergraph with $n$ vertices and $M$ hyperedges. In the critical window $M=\frac{n}{d(d-1)}(1+λn^{-1/3})+O(1)$, its largest component $L_n$ satisfies $$n^{-2/3}L_n\xrightarrow{\mathcal L} \mathcal{F}_{\mathrm{Pit}}\bigl((d-1)^{1/3}a,λ(d-1)^{-2/3}\bigr).$$ (Here $\mathcal{F}_{\mathrm{Pit}}(x,μ)$ denotes the limiting law of the largest component of the critical random graph established by Pittel (2001).) The proof rests on two pillars. First, a bivariate exponential formula gives the exact finite-size law of $L_n$. Subsequently, cancellation of the orders $n,n^{2/3},n^{1/3}$ in the tree normalization determines the series of rooted hypertrees, then that of unicyclic components and its singular expansions at the radius of convergence $ρ_d$. These calculations are exact; they serve the fixed-excess identities and the cancellation mechanism, and nothing more: no unproved uniformity in the excess enters the proof of the limit law. Next, by exploration and martingale approximation, the ordered sizes converge unconditionally to the excursion lengths of a Brownian motion with parabolic drift. A monotone coupling transfers the law to $H_d(n,M)$. The limit law is then identified with Pittel's law by a Brownian rescaling: if $h=d-1$, it suffices to multiply lengths by $h^{-1/3}$ and evaluate at parameter $λh^{-2/3}$. The exact Poisson representation dispenses with any separate deconditioning. We also formulate a two-regime $d$-uniform Bollobas bound (Wright for fixed $\ell$, global combinatorial beyond that), and explicitly indicate why neither regime alone suffices to close the analytic route.

Christoph, Müyesser and Wigderson recently asked whether an approximate form of the Erdős-Sós conjecture is robust under random edge deletions. We establish a density-sensitive transference theorem that converts global resilience for bounded-degree trees in sparse random graphs into resilient forest universality in arbitrary dense host graphs. More precisely, if $F$ is an $N$-vertex graph of edge density $λ$ bounded away from zero and $p\in[K/N,1]$, then, with probability $1-o(1)$ uniformly over the host and the percolation parameter, every subgraph obtained from $F_p$ by deleting at most an $α$-fraction of its edges contains every bounded-degree forest on at most $((1-α)λ-ξ)N$ vertices. Consequently, for every $c>0$, $L\ge1$, fixed $D$ and $α<c$, every graph $F$ on at most $Ld$ vertices with average degree at least $d$ has the property that, after percolation at any rate $p\in[K/d,1]$ and any subsequent deletion of at most an $α$-fraction of the surviving edges, the remaining graph is universal for all forests on at most $(1-c)d$ vertices and maximum degree at most $D$. This includes vertex-disjoint packings of any prescribed collection of bounded-degree trees of that total order. We further obtain forest-universality results for graphs whose connected components have vertex-cover number at most $Cd$, and for graphs that can be brought into this form by deleting sufficiently few edges on the $dn$-scale.

In 2016, Kirkland and Zeng initiated the mathematical treatment of Braess paradox for random walks on graphs using Kemeny's constant as the measure of travel time. We study this phenomenon for the Erdős-Rényi random graph $G(n,p)$. An edge in a connected graph is called Kemeny-Braess if its deletion preserves connectedness and decreases Kemeny's constant, while a nonedge is Kemeny-Braess if its insertion increases Kemeny's constant. Let $\bar d=(n-1)p$ be the expected degree of $G\sim G(n,p)$, and assume that $\bar d-\log n\to\infty$, so that $G$ is connected with high probability. We determine the sharp asymptotic thresholds for the existence of both types of paradoxical behaviour. With $C^*=1/(1-\log 2)$, the threshold for Kemeny-Braess edges is $\bar d=C^*\log n$, whereas the threshold for Kemeny-Braess nonedges is $\bar d=2C^*\log n$. Hence, as the expected degree $\bar d$ increases, the connected regime exhibits three phases: first Kemeny-Braess paradox occurs for both edges and nonedges, then only for nonedges, and finally it disappears. In particular, in the dense regime, with high probability $G$ has no Kemeny-Braess edges or nonedges. This contrasts sharply with the spectral gap of the normalised Laplacian, for which Braess' paradox persists at constant edge density. Our proof combines a spectral-moment expansion of Kemeny's constant with uniform edge-perturbation estimates. Near the thresholds, the sign of the perturbation is governed by the endpoint degrees, and binomial large deviations yield the constant $C^*$ and the factor of two between the two thresholds.

We extend Wright's enumeration of connected labelled graphs with fixed surplus to irreducible critical finite type kernels and regular compact type kernels. Let $C_t(k,r)$ denote the total weight of connected labelled graphs with $k$ vertices and surplus $r$. Let $u>0$ satisfy $Ku=u$ and $\int u\,dμ=1$, and put $α=\int u^2\,dμ$ and $β=\int u^3\,dμ$. We prove that the rooted-tree radius $ρ$ is positive and real analytic near $1$, with $ρ(1)=1$ and $\logρ(t)=\frac{α^2}{2β}(t-1)^2+O(|t-1|^3)$, and that, for fixed $r\in\mathbb{Z}_{\ge0}$ and $0\le L < \infty$, $C_t(k,r)=w_rβ^{(r-1)/2}e^{-k}ρ(t)^{-k}k^{k+3r/2-2}\{1+O_{r,L}(k^{-1/2})\}$, uniformly for $|t-1|\le Lk^{-1/2}$, where $w_r$ is the classical Wright constant. At $t_n=1+λn^{-1/3}$, these asymptotics yield local size and surplus point-process limits for deterministic finite type models and annealed compact type Bollobas-Janson-Riordan models satisfying projective quadratic regularity. After rescaling component sizes by $β^{1/3}n^{-2/3}$, the limit is the Aldous process at parameter $αβ^{-2/3}λ$, and the corresponding type profiles concentrate at the Perron law $p(dx)=u(x)μ(dx)$. For every regular countable critical kernel, we construct graph sequences with growing finite supports and the same local limits, and identify the additive shift of the critical window caused by spectral perturbations of order $n^{-1/3}$.

Consider the bipartite Erdős--Rényi graph $\mathbb{G}(n,m,d/\sqrt{mn})$ at the critical sparsity $d = b\log(N)$ for some $b > 0$ with $N = n + m$, and constant aspect ratio $γ= n/m \geq 1$. We show that the spectrum of its centered and rescaled bi-adjacency matrix, whose limiting law is the Marčenko--Pastur (MP) distribution, is split into two phases: in the bulk, singular vectors are completely delocalized, while near the edges, singular vectors are semilocalized, with their mass concentrating on small disjoint balls, centered at vertices with extreme degrees. As $b$ decreases, we identify sharp thresholds for the emergence of outlier singular values. For explicit $b_{\rm R}^{\star}$ and $b_{\rm L}^{\star}$ as functions of $γ$, we prove that when $b > b_{\rm R}^{\star}$, there is no outlier outside the bulk; when $b_{\rm L}^{\star}< b < b_{\rm R}^{\star}$, only right outliers emerge; and when $b < b_{\rm L}^{\star}$, outliers are present on both sides, all with high probability. This removes an extra technical condition from our previous work arxiv:2508.01456, providing a complete characterization of the parameter space. In sparser regimes, in terms of left outliers, we provide some nontrivial generating mechanisms, demonstrating that local graph structure plays a key role in understanding extreme singular values. Our proofs rely on two technical inputs: (1) an entrywise local law in the MP bulk, where the deterministic equivalent of the resolvent depends on both variance profiles and normalized vertex degrees, different from the Gaussian setting; (2) spectral radius upper bounds for nonbacktracking matrices constructed from regularized random graphs, recently developed in arXiv: 2609.26615. Our results reveal fruitful phenomena for sparse random rectangular matrices at criticality, which are beyond the universality class of their Gaussian counterparts.

Many algorithms are designed to work well on average over inputs. But when such an algorithm is run on an arbitrary input, the question arises: Can we trust the algorithm on this input? This question is especially relevant in network science, where algorithms for theoretical random graph models are often applied to individual real-world networks. We identify a new class of algorithmic problems addressing this, which we call "Quality Control Problems." These problems are specified by a "quality function" $ρ$ and a distribution $D$ such that, with high probability, a sample drawn from $D$ is "high quality," meaning its $ρ$-value is near $1$. The goal is to accept most inputs $x \sim D$ and reject potentially adversarially generated inputs $x$ with $ρ(x)$ far from $1$. In this work, we consider the sublinear version of the quality control problem, where $D \in Δ(\{0,1\}^N)$ and the goal is to solve the $(D,ρ)$-quality problem with $o(N)$ queries and time. As a case study, we consider random graphs, i.e., $D = G_{n,p}$ (and $N = \binom{n}2$), and the $k$-clique count function $ρ_k := C_k(G)/\mathbb{E}_{G' \sim G_{n,p}}[C_k(G')]$, where $C_k(G)$ is the number of $k$-cliques in $G$. In suitable parameter ranges, worst-case approximate $k$-clique counting requires $p^{-Ω(k^2)}$ queries. In contrast, we show that the quality control problem for $G_{n,p}$ (with $n\geq p^{-ck}$ for some constant $c$) with respect to $ρ_k$ can be tested with $p^{-O(k)}$ queries and time, showing quality control is provably superpolynomially more efficient in this setting. More generally, for a motif $H$ of maximum degree $Δ(H)$, we show that the corresponding quality control problem can be solved with $p^{-O(Δ(H))}$ queries and running time.

This paper studies crossing numbers of $d$-dimensional simplicial complexes drawn in $\mathbb R^{2d}$ or $\mathbb S^{2d}$. We clarify and generalize a number of known results and conjectures from the graph case to higher dimensions: the crossing lemma (including the exceptional case of two-dimensional complexes drawn in dimension $4$), a generalized Harary-Hill conjecture (and evidence for it), existence of a midrange crossing constant (in a weak sense), and upper bound constructions based on stochastic geometry (related to Sylvester's problem) and the cyclic polytope. Some of our results are new even in the case of graphs: we give a simple proof of the Harary-Hill conjecture for antipodal drawings, and sharpen the threshold above which the rectilinear and spherical crossing numbers of the Erdős-Rényi random graph behave like those of the complete graph.

The shortest path multiplicity, here denoted by $μ$, is an important metric of complex networks. For real networks $μ$ is high and it correlates with the network community structure. Since local network evolution induces network communities, it is possible that a high shortest path multiplicity is the natural expectation of local evolution rules. Here I demonstrate, by means of numerical simulations and exact results, that this is indeed the case. For random graphs with arbitrary degree distributions $p_k$, $\langleμ\rangle\sim \langle k(k-1)\rangle / (\langle k\rangle e)$, growing with the network size when $p_k\sim k^{-γ}$ and $γ\leq3$. For networks generated by stochastic local rules, $\langleμ\rangle$ increases with the network size and it does so faster than what is observed in their randomized versions. Furthermore, the number of communities increases with the network size and the correlation with $\langle μ\rangle$ follows.

Let $u_k^{(r)}(H,p)$ be the maximum, over all $k$-vertex $r$-uniform hypergraphs $F$, of the absolute difference between the number of induced copies of $F$ in $H$ and its expectation in the binomial random $r$-graph with the same number of vertices as $H$ and edge probability $p$. Let $u_k^{(r)}(n,p)$ be the minimum of $u_k^{(r)}(H,p)$ over all $n$-vertex $r$-uniform hypergraphs, and let $D_r(n,p)$ be the distance from $p\binom nr$ to the nearest integer. For fixed $r\ge3$, $k\ge2r$ and $\varepsilon>0$, the quantity $u_k^{(r)}(n,p)$ has order $\max\!\left\{D_r(n,p),p(1-p)\right\}n^{k-r}$ whenever $p(1-p)n^{\lfloor r/2\rfloor/(\binom{2r-2}{r}-1)}\ge\varepsilon$. Using local moment estimates and switchings in a sparse reservoir, we construct one hypergraph with the nearest-integer edge count that controls all induced orders simultaneously.

Haemers' conjecture for the generalized spectrum asserts that almost all graphs are determined by their generalized spectrum, that is, for almost all graphs $G$, every graph with the same spectrum as $G$ whose complement has the same spectrum as the complement of $G$ is isomorphic to $G$. We prove this conjecture by showing that the random graph $G(n,1/2)$ is determined by its generalized spectrum with probability tending to one. The proof combines a local theory of the denominators of the rational orthogonal matrices that relate generalized cospectral graphs, Fourier analysis of random symmetric matrices over finite rings, Seidel switching, inverse Littlewood--Offord theory over finite fields, rank estimates for random symmetric matrices modulo primes, and a study of the rational factors of the characteristic polynomial of a random symmetric $\pm1$ matrix with zero diagonal.

We prove a variational formula for the logarithmic asymptotics of a non-existence probability in a broad class of combinatorial problems such as avoiding cliques in random graphs and $k$-term arithmetic progressions in random subsets of integers. These results follow from a formula for the probability that a binomial random subset of the vertices of a locally sparse hypergraph is an independent set. The formula holds throughout the critical regime, interpolating between the regimes in which Janson's inequality and the method of hypergraph containers give the respective asymptotics. The formula is the replica-symmetric Bethe free energy formula from statistical physics applied to a hypergraph hardcore model. The proof uses two new techniques. The first, for the upper bound, involves revealing a small `window' of a random independent set and then applying entropy methods. The second, for the lower bound, involves the analysis of a random greedy algorithm guided by a Belief Propagation fixed point. In the case of avoiding cliques in random graphs, the variational formula can be expressed as an optimization problem over graphons. Moreover, typical random graphs conditioned on not containing $K_r$, when suitably normalized, approach the set of optimizing graphons in cut distance.

We introduce general methods to derive lower tail large deviation principles, for a variety of nonlinear problems of combinatorial interest. These methods are especially effective in "critical" regimes, where the nonlinearities are not sparse enough for Poissonian lower tails (e.g., Janson's inequality does not provide a sharp tail bound), but the situation is not so dense that combinatorial effects dominate (e.g., we cannot directly apply the theory of hypergraph containers or the relative entropy framework of Kozma-Samotij). In particular, these are regimes where one expects interesting phase transitions to occur. Our results have a number of consequences related to subgraphs in random graphs, answering various questions of Warnke and Jenssen-Perkins-Potukuchi-Simkin. For example, consider a random graph $G\sim \mathbb G(n,p)$, and let $L_\triangle(n,p)=\log\Pr[G\text{ is triangle-free}]$. The first-order asymptotics of $L_\triangle(n,p)$ have long been known in all parameter ranges except the critical regime where $p$ has order of magnitude $1/\sqrt n$. We are able to fill this gap: for any fixed $c>0$, we show that $n^{-3/2}L_\triangle(n,c/\sqrt n)$ converges to a limit expressed in terms of a two-parameter variational problem, which has a single phase transition at $c\approx 4.341$. In contrast, we show that there is no such phase transition for the probability that $G$ is $C_{2\ell}$-free, for any fixed even cycle $C_{2\ell}$. We also obtain some applications in combinatorial design theory. Addressing conjectures of Glock-Kühn-Lo-Osthus, Kelly, and Kwan-Sah-Sawhney-Simkin, we obtain new estimates on the number of order-$n$ Steiner triple systems with no Pasch configuration and the number of order-$n$ Latin squares with no $2\times 2$ Latin subsquare (both tight up to a factor of $\exp(o(n^2))$).

A seminal open question of Pisier and Mendel--Naor asks whether every degree-regular graph which satisfies the classical discrete Poincaré inequality for scalar functions, also satisfies an analogous inequality for functions taking values in \textit{any} normed space with non-trivial cotype. Motivated by applications, it is also greatly important to quantify the dependence of the corresponding optimal Poincaré constant on the cotype $q$. Works of Odell--Schlumprecht (1994), Ozawa (2004), and Naor (2014) make substantial progress on the former question by providing a positive answer for normed spaces which also have an unconditional basis, in addition to finite cotype. However, little is known in the way of quantitative estimates: the mentioned results imply a bound on the Poincaré constant depending super-exponentially on $q$. We introduce a novel combinatorial framework for proving quantitative nonlinear spectral gap estimates. The centerpiece is a property of regular graphs that we call \emph{long range expansion}, which holds with high probability for random regular graphs. Our main result is that any regular graph with the long-range expansion property satisfies a discrete Poincaré inequality for any normed space with an unconditional basis and cotype $q$, with a Poincaré constant that depends \emph{polynomially} on $q$, which is optimal. As an application, any normed space with an unconditional basis which admits a low distortion embedding of an $n$-vertex random regular graph, must have cotype at least polylogarithmic in $n$. This extends a celebrated lower-bound of Matoušek for low distortion embeddings of random graphs into $\ell_q$ spaces.

We study the eigenvalues of the random Laplacian matrix $D - A$, where $A$ is a Wigner matrix with sub-Gaussian entries and the diagonal matrix $D$ contains the row sums of $A$. Our main results show that the extreme eigenvalues of this model exhibit Poisson statistics; in particular, after the appropriate centering and scaling, the largest eigenvalue converges to the Gumbel distribution as the dimension of the matrix tends to infinity. This confirms, for general sub-Gaussian entries, a phenomenon the authors previously established only in the Gaussian case [Electron. J. Probab. 30 (2025), Paper No. 104], resolving a conjecture raised there. As a corollary, for an Erdős--Rényi random graph, we show the asymptotic fluctuations of the algebraic connectivity (Fiedler value) can be described in terms of the Gumbel distribution.

We study the proper hat guessing game on graphs, introduced by Adriaensen et al. in Hat guessing with proper colorings. In this game, the players are seated on the vertices of a graph $G$ and assigned hats from a set of $k$ colors such that the resulting assignment forms a proper coloring. The visibility of each vertex is limited to the hat colors of their neighborhood. Then they must simultaneously output a guess about the color of their own hat. The players win if at least one guess is correct. A parameter related to this problem is the proper hat guessing number $\operatorname{HG}_{P}(G)$ that is the maximum number of colors $m$ such that the players can guarantee a winning strategy. Motivated by the work of Shurman et al. \cite{shurman2026upper}, we establish the first upper bound that depends both on the number of vertices $n$ and the maximum degree $Δ$ in the case where $Δ\geq \frac{n}{e+1}$. This result leads us to show that the proper hat guessing number of the binomial random graph $G_{n,1/2}$ is bounded above by $cn$, where $c \approx 1.366$. Finally, we prove that graphs of maximum degree $(1-γ)n$ for some fixed $γ\in (0,1]$ cannot have $\operatorname{HG}_{P}(G) = (2-o(1))n$.