Ore-Degree in Graph Theory

• Ore-Degree is a set of degree-sum parameters that assess local constraints by summing vertex degrees over edges or nonadjacent pairs, adaptable across various graph models.
• It underpins key results such as Hamiltonicity and spanning structures by providing precise thresholds like σ₂(G) ≥ n and sharp extremal conditions.
• Extensions to oriented graphs, hypergraphs, and rainbow families reveal its flexibility, driving proofs through structural decompositions and exact threshold analysis.

Ore-degree denotes a family of degree-sum parameters derived from Ore’s theorem, rather than a single universally fixed invariant. In one common graph-theoretic usage, the Ore-degree of a graph is the maximum of $d(x)+d(y)$ over edges $xy$; in another, Ore-type conditions are expressed through the minimum degree sum over nonadjacent pairs. Later work extends the same idea to oriented graphs, hypergraphs, shadow graphs of hypergraphs, and rainbow graph families on a common vertex set, so the term is best understood as a degree-sum paradigm whose exact definition depends on the ambient category (Csaba et al., 2017, Li et al., 6 Apr 2025, Chang et al., 6 Jul 2025, Balogh et al., 6 Mar 2026).

1. Terminology and principal definitions

The literature uses several closely related Ore-type parameters. The following table records the main variants that have become standard.

Context	Parameter	Definition
Undirected graphs, edge-based	$\theta(G)$ or $\sigma(G)$	$\max_{xy\in E(G)}(\deg(x)+\deg(y))$
Undirected graphs, nonedge-based	$\sigma_2(G)$	$\min\{d_G(u)+d_G(v):u\nsim v\}$
Oriented graphs, missing arc	directed Ore threshold	$\min\{\deg^+(x)+\deg^-(y):xy\notin E(G)\}$
$r$-uniform hypergraphs	$\sigma_r(\mathcal H)$	$xy$0

For embedding problems, the edge-based form is explicit: the Ore-degree of an edge $xy$1 is $xy$2, and the Ore-degree of a graph $xy$3 is $xy$4 (Csaba et al., 2017). For Hamiltonicity, pancyclicity, and related spanning questions, the dominant parameter is instead $xy$5, the minimum degree sum over nonadjacent pairs (Li et al., 6 Apr 2025). In oriented graphs, the natural analogue for a missing directed edge $xy$6 is $xy$7, reflecting the directional asymmetry of Hamilton cycles in digraphs (Chang et al., 6 Jul 2025). In $xy$8-uniform hypergraphs, the corresponding Ore-degree is taken over non-edge $xy$9-sets and sums the degrees of all $\theta(G)$0 vertices in the set (Balogh et al., 6 Mar 2026).

A persistent source of confusion is the assumption that “Ore-degree” always means one of these formulas. The papers show instead that the phrase is contextual: some authors use it for an edge-maximum, others for a nonedge-minimum, and still others for a directed or hypergraph generalization. The unifying idea is always the same: a local obstruction is measured by a degree sum rather than by a single vertex degree.

2. Nonadjacent-pair Ore conditions and spanning structure

The classical Ore framework concerns nonadjacent vertices. In the Hamiltonian setting, one formulation is Ore’s theorem $\theta(G)$1 is hamiltonian, together with the Hamiltonian-connected analogue $\theta(G)$2 is hamiltonian-connected (Li et al., 6 Apr 2025). A path version also appears in later work: if every nonadjacent pair satisfies $\theta(G)$3, then the graph contains a Hamiltonian path (Rivera-Campo, 2012). These results motivate a large family of “Ore-type” theorems in which the threshold $\theta(G)$4 is replaced by a more refined structural quantity.

A notable refinement replaces the order $\theta(G)$5 by the bipartite-hole-number $\theta(G)$6. If $\theta(G)$7 is $\theta(G)$8-connected and $\theta(G)$9, then $\sigma(G)$0 is hamiltonian; if $\sigma(G)$1 is $\sigma(G)$2-connected and $\sigma(G)$3, then $\sigma(G)$4 is hamiltonian-connected (Li et al., 6 Apr 2025). Here $\sigma(G)$5 measures the largest size of a forced bipartite hole across all splittings, so the Ore threshold becomes structural rather than purely order-based.

The same nonadjacent-pair template also governs sparse spanning trees. For a $\sigma(G)$6-connected graph $\sigma(G)$7 with $\sigma(G)$8 and prescribed bounds $\sigma(G)$9, Rivera-Campo proved that

$\max_{xy\in E(G)}(\deg(x)+\deg(y))$0

for every nonadjacent $\max_{xy\in E(G)}(\deg(x)+\deg(y))$1 guarantees a spanning tree $\max_{xy\in E(G)}(\deg(x)+\deg(y))$2 with at most

$\max_{xy\in E(G)}(\deg(x)+\deg(y))$3

leaves and $\max_{xy\in E(G)}(\deg(x)+\deg(y))$4 for all $\max_{xy\in E(G)}(\deg(x)+\deg(y))$5 (Rivera-Campo, 2012). This theorem interpolates between Ore’s Hamilton-path condition and bounded-degree spanning tree problems.

Ore-type conditions also interact with additional global hypotheses. For $\max_{xy\in E(G)}(\deg(x)+\deg(y))$6-tough graphs, the condition

$\max_{xy\in E(G)}(\deg(x)+\deg(y))$7

forces Hamiltonicity for every $\max_{xy\in E(G)}(\deg(x)+\deg(y))$8 (Shan, 2021). For the square of a Hamiltonian cycle, an asymptotically exact Ore threshold is

$\max_{xy\in E(G)}(\deg(x)+\deg(y))$9

which implies $\sigma_2(G)$0 for all sufficiently large $\sigma_2(G)$1 (DeBiasio et al., 2014). For bootstrap percolation with threshold $\sigma_2(G)$2, the weakened condition $\sigma_2(G)$3 forces $\sigma_2(G)$4 except for explicit exceptional families $\sigma_2(G)$5 and a finite exceptional set $\sigma_2(G)$6 (Dairyko et al., 2016). In each case, Ore-type degree sums are being used as a surrogate for local expansion.

3. Edge-based Ore-degree as a graph invariant

A second major usage defines Ore-degree on edges rather than nonedges. In this sense,

$\sigma_2(G)$7

This parameter is structurally stricter than a maximum-degree bound because it controls which high-degree vertices may be adjacent (Csaba et al., 2017).

The paper “Embedding graphs having Ore-degree at most five” gives the clearest illustration. For sufficiently large $\sigma_2(G)$8, every $\sigma_2(G)$9-vertex graph $\min\{d_G(u)+d_G(v):u\nsim v\}$0 with $\min\{d_G(u)+d_G(v):u\nsim v\}$1 embeds into every $\min\{d_G(u)+d_G(v):u\nsim v\}$2-vertex graph $\min\{d_G(u)+d_G(v):u\nsim v\}$3 with minimum degree at least $\min\{d_G(u)+d_G(v):u\nsim v\}$4 (Csaba et al., 2017). The restriction $\min\{d_G(u)+d_G(v):u\nsim v\}$5 does not forbid degree-$\min\{d_G(u)+d_G(v):u\nsim v\}$6 vertices outright, but it forces them to be adjacent only to degree-$\min\{d_G(u)+d_G(v):u\nsim v\}$7 vertices. The same paper shows that when $\min\{d_G(u)+d_G(v):u\nsim v\}$8, there exists an independent dominating set $\min\{d_G(u)+d_G(v):u\nsim v\}$9 with $\min\{\deg^+(x)+\deg^-(y):xy\notin E(G)\}$0, every $\min\{\deg^+(x)+\deg^-(y):xy\notin E(G)\}$1 has $\min\{\deg^+(x)+\deg^-(y):xy\notin E(G)\}$2, and the connected components of $\min\{\deg^+(x)+\deg^-(y):xy\notin E(G)\}$3 are paths of length at most $\min\{\deg^+(x)+\deg^-(y):xy\notin E(G)\}$4 (Csaba et al., 2017). This decomposition is the central structural reason the threshold $\min\{\deg^+(x)+\deg^-(y):xy\notin E(G)\}$5 is tractable.

Edge-based Ore-degree also appears in graph coloring and criticality. For 4-critical graphs, the graphs of Ore-degree at most $\min\{\deg^+(x)+\deg^-(y):xy\notin E(G)\}$6 are exactly the 4-Ore graphs, that is, the graphs obtained from $\min\{\deg^+(x)+\deg^-(y):xy\notin E(G)\}$7 by repeated Ore-compositions (Postle, 2014). In strong edge-coloring, sparse graphs with Ore-degree $\min\{\deg^+(x)+\deg^-(y):xy\notin E(G)\}$8 and $\min\{\deg^+(x)+\deg^-(y):xy\notin E(G)\}$9 satisfy $r$0, while Ore-degree $r$1 together with $r$2 yields $r$3 (Wang, 27 Jan 2026). In the planar setting, every planar graph with $r$4 has strong chromatic index at most $r$5, settling the Chen–Huang–Yu–Zhou conjecture in the planar case (Nelson et al., 8 Sep 2025).

The parameter is also natural for line graphs. If

$r$6

then

$r$7

and in bipartite graphs the stronger bound $r$8 holds (Faron et al., 2017). Here the edge-based Ore-degree aligns directly with the local degree of a vertex in the line graph, which explains its effectiveness.

4. Directed, coloured, and hypergraph extensions

Ore-type ideas extend naturally beyond simple undirected graphs. For oriented graphs, the exact asymptotic threshold for a Hamilton cycle is: if $r$9 is an oriented graph of sufficiently large order $\sigma_r(\mathcal H)$0 and

$\sigma_r(\mathcal H)$1

then $\sigma_r(\mathcal H)$2 contains a Hamilton cycle (Chang et al., 6 Jul 2025). The bound is best possible, and the extremal construction uses a four-part orientation $\sigma_r(\mathcal H)$3 with $\sigma_r(\mathcal H)$4 (Chang et al., 6 Jul 2025). A different oriented variant studies discrepancy rather than mere existence: if an oriented graph satisfies $\sigma_r(\mathcal H)$5, then it contains a Hamilton cycle $\sigma_r(\mathcal H)$6 with

$\sigma_r(\mathcal H)$7

for every $\sigma_r(\mathcal H)$8 and sufficiently large $\sigma_r(\mathcal H)$9 (Chang et al., 19 Mar 2026).

In rainbow graph families, the Ore-type parameter becomes family-valued. For a family $xy$00 on a common $xy$01-vertex set,

$xy$02

If $xy$03, then either $xy$04 is $xy$05-rainbow vertex-pancyclic or $xy$06; if

$xy$07

then $xy$08 is rainbow vertex-pancyclic (Li et al., 30 Apr 2026). This is an Ore-type family analogue of Bondy’s pancyclicity theorem.

Hypergraph versions bifurcate according to the Hamiltonicity model. For $xy$09-graphs, the relevant condition is imposed on the $xy$10-shadow $xy$11: there exists a constant $xy$12 such that if every nonadjacent pair $xy$13 in $xy$14 satisfies

$xy$15

then $xy$16 contains a Hamiltonian Berge cycle, with the proved value $xy$17 and conjectured optimal value $xy$18 (Li et al., 17 May 2025). For $xy$19-uniform hypergraphs, the Ore-degree is

$xy$20

and if

$xy$21

then $xy$22 contains $xy$23 pairwise disjoint edges, provided $xy$24 and $xy$25 (Balogh et al., 6 Mar 2026). These extensions show that Ore-type degree sums can be transferred from edges and nonedges to shadow adjacency, missing arcs, and non-edge $xy$26-sets.

5. Structural role and proof methods

Ore-degree conditions are typically not used in isolation; they are converted into overlap statements on neighborhoods, rigid decompositions, or reducible configurations. In bounded-degree spanning tree theory, the Ore hypothesis is fed into a maximal constrained subtree argument, combined with Menger’s theorem and edge exchanges that reduce the number of leaves (Rivera-Campo, 2012). In the bipartite-hole-number Hamiltonicity theorem, the proof proceeds through a maximal-edge counterexample, Hamilton-path rerouting, and forbidden bipartite-hole counting (Li et al., 6 Apr 2025).

In dense embedding theory, Ore-degree bounds often serve as the structural input to regularity methods. For $xy$27, the spanning embedding theorem uses Szemerédi regularity, triangle factors in the reduced graph, proportional and strong proportional matchings, Csaba’s modified Blow-up Lemma, and an extremal analysis around triangle-rich targets (Csaba et al., 2017). For the exact oriented Hamiltonicity threshold, the proof combines an absorbing path, a reservoir lemma, a reduced oriented graph inheriting an approximate Ore condition, and a stability analysis of near-extremal four-part structures (Chang et al., 6 Jul 2025). In oriented discrepancy, a two-step absorption method is paired with a reduced-graph Ore bound and an Ore-type Hajnal–Szemerédi-style tournament tiling (Chang et al., 19 Mar 2026).

In sparse graph coloring, Ore-degree restrictions drive discharging. The strong edge-coloring results for Ore-degree $xy$28 or $xy$29 use minimal counterexamples, partial colorings, Hall’s marriage theorem, and carefully tuned discharging rules indexed by local degree types (Wang, 27 Jan 2026). For planar graphs with $xy$30, the proof adds reducible configurations certified by Combinatorial Nullstellensatz to a discharging argument based on the fact that only degrees $xy$31 can occur, that every $xy$32-vertex is adjacent to two $xy$33-vertices, and that short faces have tightly constrained degree patterns (Nelson et al., 8 Sep 2025).

A common pattern across these arguments is that the degree-sum hypothesis collapses many potential local configurations. In edge-based settings, it suppresses adjacency among high-degree vertices; in nonedge-based settings, it forces overlap between structured neighborhood sets. This suggests that Ore-degree is best viewed as a local incompatibility bound: it excludes sparse separators in some problems and excludes dense local clashes in others.

6. Extremality, sharpness, and scope

A striking feature of Ore-degree theory is the frequency of exact or near-exact thresholds. Rivera-Campo’s spanning-tree theorem is sharp: a complete bipartite graph $xy$34 with $xy$35 and $xy$36 meets the degree-sum threshold minus one, yet every spanning tree violates at least one prescribed degree bound (Rivera-Campo, 2012). In the toughness setting, the conjectural bound

$xy$37

would be best possible if true, with complete bipartite and join constructions furnishing equality examples (Shan, 2021).

The exact oriented Hamiltonicity threshold $xy$38 is also best possible: the four-part construction $xy$39 with $xy$40 has no Hamilton cycle, yet one can tune the sizes and internal tournaments so that some missing arc $xy$41 satisfies

$xy$42

(Chang et al., 6 Jul 2025). In rainbow pancyclicity, the unique obstruction at threshold $xy$43 is the family $xy$44, exactly paralleling Bondy’s classical extremal graph (Li et al., 30 Apr 2026). In oriented discrepancy, blow-ups of transitive tournaments show that the coefficient $xy$45 in the lower bound $xy$46 cannot be improved (Chang et al., 19 Mar 2026). For bootstrap percolation, the threshold $xy$47 is accompanied by explicit infinite exceptional classes $xy$48 and a finite set $xy$49 (Dairyko et al., 2016).

The scope of the term also has a negative aspect: it should not be conflated with degree in the algebraic theory of Ore extensions or Ore operators. In work on graded iterated Ore extensions and Ore operators, “degree” refers to internal grading, coefficient degree, or order-degree tradeoffs, not to graph-theoretic degree sums (Elle, 2015, Chen et al., 2013). In combinatorics proper, however, Ore-degree has become a flexible language for expressing local density through degree sums, and its most successful applications occur precisely where such sums reveal a hidden structural rigidity.