Alphabeta Math
How statement and proof provenance work

The first chip identifies the source of the statement or construction; the second identifies the source of its local proof or verification.

  • Literature-sourced: the exact statement appears in a cited source; only wording and notation differ.
  • AI-adapted: a semantically identical restatement of literature-sourced material, modulo indexing, notation, and boundary cases adopted by the library.
  • AI-generated: a genuinely novel statement formulated by AI, with no source for the claim itself.

These labels describe origin, not correctness: citations and verification chips remain separate evidence.

✓ 6 results · all verified · 4 also independently AI-judged
Every result on this page is machine-checked by a proof checker and read in full and owner-audited; the judge is an additional, independent cross-model AI review of the proofs. The 2 not AI-judged were verified by owner audit (typically over a confirmed judge false positive), not failures.

From Generalized Niceness to Erdős-Hajnal

1 · Prerequisites

2 · Summary

This page closes the generalized-niceness route. The preceding page produced a constant-scale restricted theorem; the present page adds the Rödl initialization that removes the starting restriction and then isolates the blockade hypothesis needed for the published blockade-to-restricted theorem.

The final theorem converts those local restricted or blockade outcomes into a polynomial clique or stable set and then invokes complement invariance to pass from the F‾-free formulation used in generalized niceness back to the ordinary Erdős-Hajnal property of F.

3 · Logical flowchart

4 · Definitions, theorems and proofs

LemmaStatement: Literature-sourcedProof: AI-generatedprecheck passaudited 2026-09-05Open item page →

Rödl initialization upgrades generalized niceness to a restricted set, a complete or anticomplete blockade, or a polynomial clique or stable set

Statement

Let F be a generalized nice, leaf-reducible, wonderful finite family. Then there exist constants c1≥1 and c2>0 such that for every x∈(0,12) and every F‾-free graph G, at least one of the following holds:

  1. G has an x-restricted induced subgraph with at least xc1∣G∣ vertices;
  2. G has a complete or anticomplete (k,∣G∣/kc1)-blockade for some integer k≥2;
  3. G has a clique or stable set of size at least (xc1∣G∣)c2.

Facts & Assumptions

Given: A generalized nice, leaf-reducible, wonderful finite family F, a parameter x∈(0,12), and an F‾-free graph G.

[L1]

There exist constants c∈(0,12), a1≥1, and a2>0 such that every c2-restricted F‾-free graph satisfies the three-outcome conclusion with parameter x whenever x<c2 (Constant-scale restricted generalized niceness yields an x-scale restricted subgraph, a polynomial clique or stable set, or a blockade).

[L2]

For every graph H and every ξ∈(0,12), every nonempty H-free graph has a ξ-restricted induced subgraph of size at least δ∣G∣ for some constant δ>0 depending only on H and ξ (Rödl: for every H and every ϵ∈(0,12) there is δ>0 such that every nonempty H-free graph has an ϵ-restricted vertex set of size at least δ∣V(G)∣).

Proof

technique · use Rödl at the fixed scale $\xi=c^2$, then apply the constant-scale theorem unless $x$ is already above that scale
1.1L1L2choose

Let c∈(0,12), a1≥1, and a2>0 be the constants supplied by [L1], and set ξ:=c2. Choose δ>0 from [L2] for the forbidden family F‾ and the parameter ξ.

2.1step 1.1choose

Choose an integer c1≥1 so large that δ≥ξc1/2,δ≥2a1−c1,c1≥2a1, and set c2:=a2.

2.2step 1.1L2

By [L2], the graph G has a ξ-restricted induced subgraph F⊆G with ∣F∣≥δ∣G∣.

3.1step 2.1step 2.2givenalgebra

Suppose first that x≥ξ. Then F is also x-restricted, and ∣F∣≥δ∣G∣≥2a1−c1∣G∣≥xc1−a1∣G∣≥xc1∣G∣, because x≤12 and a1≥1. Hence outcome 1 holds.

3.2step 1.1step 2.2L1

Assume now that x<ξ=c2. Then [L1] applies to the c2-restricted graph F and yields one of three conclusions: an x-restricted induced subgraph S⊆F with ∣S∣≥xa1∣F∣, a clique or stable set of size at least (xa1∣F∣)a2, or a complete or anticomplete (k,∣F∣/ka1)-blockade in F for some integer k≥2.

4.1step 2.1step 3.2algebra

In the first branch, xa1∣F∣≥xa1δ∣G∣≥xa1ξc1/2∣G∣≥xa1+c1/2∣G∣≥xc1∣G∣, because x<ξ and c1≥2a1. So outcome 1 holds.

4.2step 2.1step 3.2algebra

In the third branch, ∣F∣ka1≥δ∣G∣ka1≥2a1−c1∣G∣ka1≥∣G∣kc1, because k≥2. Thus the same blocks give outcome 2 in G.

5.1step 2.1step 3.2step 4.1algebra

In the second branch, the same inequality from step 4.1 gives (xa1∣F∣)a2≥(xc1∣G∣)c2, so outcome 3 holds.

6.1step 3.1step 3.2step 4.1step 5.1step 4.2∎

Steps 3.1, 4.1, 4.2, and 5.1 cover all possibilities, so one of the three stated outcomes always holds.

LemmaStatement: Literature-sourcedProof: AI-generatedprecheck passjudge pass (gpt-5.6-terra)audited 2026-09-05Open item page →

Large induced subgraphs without a polynomial clique or stable set force complete or anticomplete blockades

Statement

Let F be a generalized nice, leaf-reducible, wonderful finite family. Let c1≥1 and c2>0 be the constants from Rödl initialization upgrades generalized niceness to a restricted set, a complete or anticomplete blockade, or a polynomial clique or stable set. Fix an F‾-free graph G, define

q:=42c12,c:=min⁡{q−1,c2/2},x:=∣G∣−1/(3c1),ϵ:=x1/(7c1).

and assume that G has no clique or stable set of size at least ∣G∣c. Then every induced subgraph F of G with ∣F∣≥ϵ2c1∣G∣ has a complete or anticomplete (k,∣F∣/kc1)-blockade for some integer k∈[2,ϵ−1].

Facts & Assumptions

Given: The data and hypotheses in the Statement.

[L1]

The previous lemma gives every F‾-free graph either an x-restricted induced subgraph of size at least xc1 times the ambient order, or a complete or anticomplete (k,∣G∣/kc1)-blockade with k≥2, or a clique or stable set of size at least (xc1∣G∣)c2 (Rödl initialization upgrades generalized niceness to a restricted set, a complete or anticomplete blockade, or a polynomial clique or stable set).

[L2]

A nonempty x-sparse graph H satisfies χ(H)≤x∣H∣+1,∣H∣≤χ(H)α(H), by The greedy colouring bound χ(G)≤Δ(G)+1 for every nonnull finite graph and The bounds ω(G)≤χ(G) and ∣V(G)∣≤χ(G)α(G).

Proof

technique · apply the previous lemma to $F$ and show that the restricted and clique/stable branches contradict the assumed failure of the polynomial bound
1.1givenassume-contra

Let F be an induced subgraph of G with ∣F∣≥ϵ2c1∣G∣, and suppose for contradiction that F has no complete or anticomplete (k,∣F∣/kc1)-blockade for any integer k∈[2,ϵ−1].

2.1

Apply [L1] to the graph F with the parameter x. Because the blockade branch is excluded by step 1.1, either:

  1. F has an x-restricted induced subgraph S with ∣S∣≥xc1∣F∣, or
  2. F has a complete or anticomplete (k,∣F∣/kc1)-blockade for some integer k≥2, or
  3. F has a clique or stable set of size at least (xc1∣F∣)c2.

[step 1.1, L1]

3.1step 2.1L3algebra

Suppose the restricted branch of step 2.1 holds. Then ∣S∣≥xc1∣F∣≥xc1ϵ2c1∣G∣=xc1+2/7∣G∣=x−2c1+2/7. Since c1≥1 and 0<x<1, the exponent −2c1+2/7 is at most −1, so ∣S∣≥x−1. After replacing S by the same set in the complementary graph if necessary, [L3] lets us assume that S is x-sparse.

3.2step 1.1step 2.1algebrachoose

Suppose instead that the blockade branch of step 2.1 holds. Then step 1.1 forces k>ϵ−1. Choosing one vertex from each block gives a clique or stable set of size k>ϵ−1=∣G∣1/(21c12)≥∣G∣c, because c≤(42c12)−1. This contradicts the hypothesis on G.

3.3step 2.1algebra

Suppose instead that the clique-or-stable-set branch of step 2.1 holds. Then (xc1∣F∣)c2≥(xc1ϵ2c1∣G∣)c2=(xc1+2/7∣G∣)c2. Since x=∣G∣−1/(3c1), the inner factor equals ∣G∣1−(c1+2/7)/(3c1), whose exponent is at least 1/2 because c1≥1. Therefore F contains a clique or stable set of size at least ∣G∣c2/2≥∣G∣c, because c≤c2/2. This again contradicts the hypothesis on G.

4.1step 3.1L2algebra

By [L2], α(G[S])≥∣S∣x∣S∣+1=1x+∣S∣−1≥1x+x≥x−1/2. Because x=∣G∣−1/(3c1), this gives a clique or stable set of size at least ∣G∣1/(6c1)≥∣G∣c, contradicting the hypothesis on G because c≤q−1=1/(42c12)≤1/(6c1).

5.1step 4.1step 3.2step 3.3discharge-contradiction∎

All three branches from step 2.1 contradict the hypothesis on G, so the assumption in step 1.1 was false. Therefore F has a complete or anticomplete (k,∣F∣/kc1)-blockade for some integer k∈[2,ϵ−1].

TheoremStatement: Literature-sourcedProof: AI-generatedprecheck passjudge pass (gpt-5.6-terra)audited 2026-09-05Open item page →

Leaf-reducible wonderful generalized nice finite families have the Erdős-Hajnal property

Statement

Let F be a generalized nice, leaf-reducible, wonderful finite family. Then F has the Erdős-Hajnal property.

Facts & Assumptions

Given: A generalized nice, leaf-reducible, wonderful finite family F.

[L1]

There exist constants c1≥1 and c2>0 such that every F‾-free graph satisfies the three-outcome lemma from Rödl initialization upgrades generalized niceness to a restricted set, a complete or anticomplete blockade, or a polynomial clique or stable set.

[L2]

For those constants, every induced subgraph F of an F‾-free graph G with ∣F∣≥ϵ2c1∣G∣ has a complete or anticomplete (k,∣F∣/kc1)-blockade for some k∈[2,ϵ−1] whenever G has no clique or stable set of size ∣G∣c, where c:=min⁡{(42c12)−1,c2/2}. (Large induced subgraphs without a polynomial clique or stable set force complete or anticomplete blockades)

[L3]

If every induced subgraph F of G with ∣F∣≥ϵ2a∣G∣ has a complete or anticomplete (k,∣F∣/ka)-blockade for some k∈[2,ϵ−1], then G has an ϵ-restricted induced subgraph with at least ϵ3a∣G∣ vertices (Complete or anticomplete blockade hypotheses force an ϵ-restricted induced subgraph).

Proof

technique · prove an Erdős-Hajnal exponent for the $\overline{\mathcal F}$-free class and then use complement invariance
1.1L1choosegiven

Let c1≥1 and c2>0 be the constants from [L1], put q:=42c12,m:=2q,c:=min⁡{q−1,c2/2}, and fix a nonempty F‾-free graph G. We will prove that G has a clique or stable set of size at least ∣G∣c.

2.1step 1.1assume-contraalgebra

If G already has such a clique or stable set, there is nothing to prove. So assume for contradiction that G has no clique or stable set of size ∣G∣c. Because c≤q−1, one has mc=(2q)c≤2. Hence every nonempty graph with at most m vertices already has a clique or stable set of size at least ∣G∣c, so this forces ∣G∣>m.

3.1step 2.1algebra

Define x:=∣G∣−1/(3c1),ϵ:=x1/(7c1). Then x≤m−1/(3c1)=2−14c1,ϵ≤2−14c1/(7c1)=14, so in particular x<ϵ≤14.

4.1step 2.1step 3.1L2L3

By [L2], every induced subgraph F of G with ∣F∣≥ϵ2c1∣G∣ has a complete or anticomplete (k,∣F∣/kc1)-blockade for some k∈[2,ϵ−1]. Therefore [L3] applies with a:=c1 and gives an ϵ-restricted induced subgraph S⊆G with ∣S∣≥ϵ3c1∣G∣.

5.1step 3.1step 4.1L4algebra

Since ϵ=x1/(7c1)=∣G∣−1/(21c12) and c1≥1, one has ∣S∣≥ϵ3c1∣G∣=ϵ−21c12+3c1≥ϵ−1. After taking complements if necessary, [L4] lets us assume that S is ϵ-sparse.

6.1step 5.1L4algebra

Applying [L4] to the graph G[S] yields α(G[S])≥∣S∣ϵ∣S∣+1=1ϵ+∣S∣−1≥12ϵ≥ϵ−1/2=∣G∣1/(42c12)≥∣G∣c, because c≤(42c12)−1. This contradicts step 2.1.

7.1step 1.1step 6.1L5discharge-contradiction∎

Therefore every nonempty F‾-free graph has a clique or stable set of size at least ∣G∣c, so the complement class of F has the Erdős-Hajnal property. By [L5], F itself has the Erdős-Hajnal property.

5 · Examples, counterexamples and false statements

ExampleConstruction: AI-generatedVerification: AI-generatedprecheck passaudited 2026-09-05Open item page →

The Lemma 3.5 parameter choice at the next power of two above the source threshold

Example

Consider the special case in which the constant supplied by the source lemma happens to be c1=1. Then q=42 and the threshold is m=242. Let ∣G∣:=243, and define

x:=∣G∣−1/3=2−43/3,ϵ:=x1/7=2−43/21.

Then the size thresholds in the source Erdős-Hajnal reduction are

∣F∣min⁡:=ϵ2∣G∣,∣S∣min⁡:=ϵ3∣G∣.

Facts & Assumptions

Given: The numerical choices displayed in the Example.

[A1]

Conditionally on the displayed value c1=1, one has ∣G∣=243>242=m, and 243 is the first power of two strictly above the corresponding source threshold. The example does not assert that the existential constant c1 can be freely chosen.

Verification

technique · direct arithmetic
1.1A1algebra

Since x=2−43/3, one has ϵ=x1/7=2−43/21>2−43/3=x. Also ϵ=2−43/21<2−2=14, because 43/21>2. Thus x<ϵ≤14.

2.1step 1.1algebra

The first threshold is ∣F∣min⁡=ϵ2∣G∣=2−86/21⋅243=2817/21, while ϵ−1=243/21. Since 817/21>43/21, one gets ∣F∣min⁡>ϵ−1.

2.2step 1.1algebra

The second threshold is ∣S∣min⁡=ϵ3∣G∣=2−129/21⋅243=2774/21. Again 774/21>43/21, so ∣S∣min⁡>ϵ−1.

3.1step 2.1step 2.2∎

Therefore, in the special case c1=1, this concrete instance satisfies all of the numerical inequalities used in the source Erdős-Hajnal reduction at the next power-of-two graph order above the threshold.

ExampleConstruction: AI-generatedVerification: AI-generatedprecheck passjudge pass (gpt-5.6-terra)audited 2026-09-05Open item page →

A complete four-blockade gives a four-vertex clique

Example

Let B=(B1,B2,B3,B4) be a complete four-blockade in a graph G, and choose vertices vi∈Bi for i=1,2,3,4.

Facts & Assumptions

Given: The complete blockade B and the chosen vertices vi∈Bi.

[L1]

Distinct blocks of a complete blockade are pairwise complete (Complete, anticomplete, pure, weakly sparse, and x-sparse blockades).

Verification

technique · direct
1.1givenchoose

Because each block of a blockade is nonempty, the four choices vi∈Bi are legitimate.

2.1step 1.1L1

If i≠j, then Bi is complete to Bj by [L1], so the chosen vertices vi and vj are adjacent.

3.1step 2.1∎

Therefore every pair among v1,v2,v3,v4 is adjacent, so these four vertices form a clique.

ExampleConstruction: AI-generatedVerification: AI-generatedprecheck passjudge pass (gpt-5.6-terra)audited 2026-09-05Open item page →

A large epsilon-restricted induced subgraph gives a polynomial clique or stable set

Example

Let ϵ:=14 and let S be an ϵ-restricted induced subgraph on 16 vertices. Take an ambient graph order ∣G∣:=284, so that ∣G∣1/42=4.

Facts & Assumptions

Given: The data in the Example.

[L2]

For a nonempty ϵ-sparse graph H, χ(H)≤ϵ∣H∣+1,∣H∣≤χ(H)α(H) (The greedy colouring bound χ(G)≤Δ(G)+1 for every nonnull finite graph, The bounds ω(G)≤χ(G) and ∣V(G)∣≤χ(G)α(G)).

Verification

technique · direct
1.1L1

By [L1], after taking complements if necessary we may assume that S is ϵ-sparse.

2.1step 1.1L2algebra

Applying [L2] to the 16-vertex graph G[S] gives α(G[S])≥16ϵ⋅16+1=165>3. Since α(G[S]) is an integer, α(G[S])≥4.

3.1step 2.1L1algebra∎

Because ∣G∣=284, one has ∣G∣1/42=22=4. Therefore the stable set from step 2.1 already has the same size as the final polynomial bound used in the A-page reduction. If the sparse side had arisen in the complement instead, [L1] would turn the same calculation into a clique of size 4 in the original graph.

Sources