Erdős Matching Conjecture Overview

• Erdős Matching Conjecture is a fundamental problem in extremal set theory that bounds the size of k-uniform families with restricted matching numbers using two canonical constructions.
• Recent progress, culminating in Mishra’s full resolution, leverages shifting techniques and stability analyses to establish the extremal bounds for all n ≥ sk.
• Its methods, including the truncated (clique) and star-like (cover) constructions, have significant applications in hypergraph matching, finite geometry, and probabilistic combinatorics.

The Erdős Matching Conjecture (EMC) is a fundamental problem in extremal set theory and hypergraph theory, concerned with bounding the size of $k$-uniform set families and hypergraphs that avoid large matchings. It asserts that, for given $n,k,s$, the largest size of a $k$-uniform family of $k$-sets over an $n$-element ground set with matching number at most $s$ is attained by one of two canonical constructions. The conjecture has spawned a vast literature, with recent progress culminating in its full resolution.

1. Formal Statement and Extremal Constructions

Let $[n]=\{1,\ldots,n\}$. A $k$-uniform family (or $k$-graph) $\mathcal{F} \subset \binom{[n]}{k}$ is a collection of $n,k,s$0-element subsets. The matching number $n,k,s$1 denotes the size of the largest collection of pairwise disjoint sets in $n,k,s$2. The Erdős Matching Conjecture states:

For all $n,k,s$3, $n,k,s$4, $n,k,s$5, every $n,k,s$6-uniform family $n,k,s$7 with $n,k,s$8 satisfies

$n,k,s$9

The two canonical extremal families are:

• Truncated Family (Clique Construction): All $k$0-sets from a fixed ground set of $k$1 elements: $k$2 (matching number $k$3).
• Star-like Family (Cover Construction): All $k$4-sets that intersect a fixed $k$5-element set $k$6: $k$7 (matching number $k$8).

Both constructions saturate the bound, and for $k$9, the star-like family yields the extremal size.

2. Historical Progress and Resolution

Erdős originally posed this problem in 1965. Early results affirmed EMC for $k$0 (Erdős–Gallai theorem) and for $k$1 asymptotically. Kleitman solved the $k$2 case, showing the truncated family is unique. In larger $k$3, significant advances were due to work by Frankl and Kupavskii, who improved the $k$4 threshold for which EMC could be proved, ultimately reducing it to nearly twice the conjectured minimum $k$5.

The full conjecture was recently resolved by Mishra (Mishra, 1 Feb 2026), who established the upper bound for all $k$6 and arbitrary $k$7. The proof leverages a stabilization process using shifting operations and potential functions, iteratively modifying the family to one of the two extremal forms without increasing its size or matching number.

For the case $k$8, the conjecture was confirmed by Frankl (Luczak et al., 2012) for all sufficiently large $k$9, using a combination of the shifting technique, stability analysis, and intricate counting of multipartite contributions. For $n$0, the recent global solution applies.

3. Methods, Stability, and Structure Theorems

A central technique in EMC proofs is shifting, an operation on set families pushing the sets toward lex-order minimality while preserving the matching number and size. The stabilization via shifts produces families that are highly structured, facilitating combinatorial or probabilistic arguments.

Stability versions, especially those of Hilton–Milner type, show that if a family achieves (or nearly achieves) the extremal size, then it must be isomorphic to one of the canonical constructions (Frankl et al., 2016, Martin et al., 2024). For large $n$1, families with maximum matching number and minimal degree close to the unique extremal have strictly smaller size unless they match the trivial structure exactly.

For certain ranges, a third, Hilton–Milner–type construction arises as the unique runner-up when the covering number exceeds the matching number, characterizing all almost-extremal families (Frankl et al., 2016).

4. Extensions, Variants, and Degree-Threshold Results

Variants and Generalizations

• $n$2-matching number: The extremal problem generalizes to families avoiding certain intersections (i.e., requiring $n$3), leading to an analogue of EMC where the extremal structures are unions of $n$4-stars (zhang et al., 18 Aug 2025, Pelekis et al., 2017).
• General $n$5-Property: Families where the union of any $n$6 sets has size at most $n$7 unify EMC with intersection theorems; explicit extremal families and thresholds are known for large $n$8 (Frankl et al., 2019).
• $n$9-analogs: Similar conjectures in finite vector spaces (subspace settings) identify extremal families as unions of "dictator" classes. For $s$0 bounded relative to $s$1 and $s$2, optimality is achieved by the union of $s$3 dictator classes (Ihringer, 2020).

Degree Thresholds

• Minimum Vertex Degree: Lower bounds on vertex degree imply large or perfect matchings in $s$4-graphs. Recent improvements lower the $s$5 threshold and identify sharp thresholds for perfect matchings in hypergraphs, using fractional matching, random sparsification, and stability techniques (Guo et al., 2020, Han, 2015).
• Perfect and Almost Perfect Matchings: New upper bounds for $s$6-degree thresholds for perfect matchings follow from EMC-type results, and are exact for a wide range of parameters ($s$7 close to $s$8) (Han, 2015).

5. Probabilistic and Random Hypergraph Analogues

The EMC has natural interpretations in the theory of Kneser hypergraphs. The independence number of the $s$9-uniform Kneser hypergraph $[n]=\{1,\ldots,n\}$0 is precisely the extremal function considered by EMC. A random version, where each edge is selected independently with probability $[n]=\{1,\ldots,n\}$1, exhibits a sharp threshold: when $[n]=\{1,\ldots,n\}$2 exceeds a critical value, the random hypergraph a.s. inherits the EMC extremal property, and maximum independent sets remain unions of stars (Alishahi et al., 2018).

This line of work connects EMC to probabilistic combinatorics, concentration inequalities, and anti-Ramsey theory.

6. Impact, Applications, and Open Directions

The resolution and partial results on the EMC have profound implications in extremal combinatorics, theoretical computer science (e.g., Dirac-type conditions, property testing), and finite geometry (Cameron–Liebler line classes). The methods developed—shifting, cross-dependent families, random chain analysis, and spectral techniques—are foundational in adjacent areas.

Open directions include:

• Precise stability and diversity theorems for near-extremal families (especially at the precise thresholds $[n]=\{1,\ldots,n\}$3 for small $[n]=\{1,\ldots,n\}$4).
• Exact degree-thresholds for perfect matchings in all parameter ranges.
• Analysis of random versions near and within the critical window.
• Extension of $[n]=\{1,\ldots,n\}$5-matching and $[n]=\{1,\ldots,n\}$6-bounds beyond the currently accessible ranges.
• Further exploration of $[n]=\{1,\ldots,n\}$7-analogs in vector spaces and related geometries.

7. Table: Key Milestones in EMC

Result or Parameter Regime	Reference	Key Contribution
$[n]=\{1,\ldots,n\}$8 (graphs)	Erdős–Gallai	Exact solution
$[n]=\{1,\ldots,n\}$9	Kleitman	Uniqueness, truncated family extremal
Large $k$0 fixed	Frankl, Kupavskii	EMC holds for $k$1 etc.
$k$2, all large $k$3	Frankl	EMC resolved (Luczak et al., 2012)
Almost-perfect matchings	Kolupaev–Kupavskii	Extremal window broadened (Kolupaev et al., 2022)
Full solution, all $k$4	Mishra	Conjecture resolved (Mishra, 1 Feb 2026)
$k$5-matching and $k$6 cases	Kupavskii, Frankl	Generalized extremal structures (zhang et al., 18 Aug 2025, Frankl et al., 2019)

The EMC stands as a central result illuminating deep combinatorial dichotomies—between clique-like and covering-like extremal structures—and underpins much of the ongoing research in extremal set theory and hypergraph matching theory.