AFTD Auto-Formalizing Theoretical Domains

wild_joint_even_canonical_orbit_obstruction

theoremverified

At q=4, the canonical triples {1,2,11} and {1,2,3} have the same declared type, but their conditional whole-bad matching probabilities are respectively 1 and strictly less than 1. Thus the helper asserting type invariance for all q>=3 is false; an odd-q hypothesis is essential. This does not refute the main conjecture or the parent mixed moment, which assumes odd q.

Statement

theorem wild_joint_even_canonical_orbit_obstruction :
    wild_joint_canonical_type 4 (by omega) {1,2,11} =
      wild_joint_canonical_type 4 (by omega) {1,2,3} ∧
    wild_joint_alpha 4 {1,2,11} = 1 ∧ wild_joint_alpha 4 {1,2,3} < 1
Source
IN-382 / OP-306 campaign9. Refutes the AFTD agent helper wild_joint_alpha_type_well_defined as stated, not a published paper claim. Discovered by exact experiment9566; Lean proves the unequal probabilities using explicit connected/disconnected certificates without evaluating 12! permutations.
Verified
30 Oct 2026
Axioms
Classical.choiceQuot.soundpropext
Built on 6
  • wild_joint_alphaalpha(T): the conditional probability, under the uniform distribution on all vertex permutations σ, that the fixed canonical matching is…
  • wild_joint_canonical_typeThe orbit type in {0,1,2,3} of a triple T relative to the fixed canonical matching: 0 (=A) if T contains a canonical edge and 0; 1 (=B) if…
  • wild_joint_bad_relationA symmetric relation R is whole-bad for the block map p if some nonempty proper union of blocks (a nonempty proper set of block indices B,…
  • wild_joint_canonical_matchingThe fixed matching on Fin (3q) with isolated vertex 0 and pairs {2k-1, 2k} for k >= 1: (1,2),(3,4),(5,6),...
  • wild_joint_first_blockThe first block: the set of vertices placed in positions 0,1,2 by σ (= {σ 0, σ 1, σ 2}).
  • wild_joint_blockBlock (index of the consecutive triple) containing vertex v under the permutation σ, where σ p is the vertex at position p; blocks are…
Used by 1

The proof, explained

For q = 4, the triples {1,2,11} and {1,2,3} both have canonical type 2 because each contains the matched edge {1,2} and neither contains 0. However, {1,2,11} is closed under the canonical matching, forcing all permutations with first block {1,2,11} to admit a bad cut with block {0}, so alpha({1,2,11}) = 1. In contrast, an explicit permutation with first block {1,2,3} is verified to not be bad, yielding alpha({1,2,3}) < 1.

  1. Define in382_good_perm and in382_bad_perm and check their inverse properties.

    The permutations on Fin 12 are defined with explicit forward and inverse lookup tables, and the inverses are verified by exhausting over all 12 elements.

    lines 8–15

      left_inv := by intro x; fin_cases x <;> decide
      right_inv := by intro x; fin_cases x <;> decide
    
    def in382_bad_perm : Equiv.Perm (Fin 12) where
      toFun := ![1,2,11,0,3,4,5,6,7,8,9,10]
      invFun := ![3,0,1,4,5,6,7,8,9,10,11,2]
      left_inv := by intro x; fin_cases x <;> decide
      right_inv := by intro x; fin_cases x <;> decide
  2. Compute first blocks for both permutations.

    Extensionality and exhaustive case analysis on Fin 12 verify that the first blocks are {1,2,3} and {1,2,11} respectively.

    lines 18–27

      classical
      ext x
      simp only [wild_joint_first_block, mem_filter, mem_univ, true_and]
      fin_cases x <;> norm_num [wild_joint_first_block,wild_joint_block,in382_good_perm, Fin.ext_iff]
    
    theorem in382_bad_first : wild_joint_first_block 4 in382_bad_perm = {1,2,11} := by
      classical
      ext x
      simp only [wild_joint_first_block, mem_filter, mem_univ, true_and]
      fin_cases x <;> norm_num [wild_joint_first_block,wild_joint_block,in382_bad_perm, Fin.ext_iff]
  3. The good permutation is not bad.

    Unfolding definitions and running kernel evaluation shows the good permutation does not satisfy the bad relation.

    lines 32–33

      unfold wild_joint_bad_relation wild_joint_block wild_joint_canonical_matching
      decide +kernel
  4. The bad triple is closed under canonical matching.

    Case analysis on elements of {1,2,11} and target y shows matching edges stay within the triple.

    lines 37–40

      intro x hx y hxy
      simp only [mem_insert,mem_singleton] at hx
      rcases hx with rfl|rfl|rfl <;> fin_cases y <;>
        norm_num [wild_joint_canonical_matching, Fin.ext_iff] at *
  5. Any permutation with first block {1,2,11} is bad.

    The singleton block {0} witnesses the bad relation because block 0 corresponds to {1,2,11}, which is closed under the matching.

    lines 45–62

      classical
      refine ⟨{0},by simp,by decide,?_⟩
      have hmem : ∀ x, wild_joint_block 4 σ x ∈ ({0} : Finset (Fin 4)) ↔
          x ∈ ({1,2,11} : Finset (Fin 12)) := by
        intro x
        rw [← hσ]
        simp only [mem_singleton, wild_joint_first_block, mem_filter, mem_univ, true_and]
        change (⟨(σ.symm x).val / 3, _⟩ : Fin 4) = 0 ↔ (σ.symm x).val < 3
        rw [Fin.ext_iff]
        simp only [Fin.val_zero, Nat.div_eq_zero_iff]
        norm_num []
      intro x y hxy
      rw [hmem, hmem]
      constructor
      · exact fun hx => in382_bad_closed x hx y hxy
      · intro hy
        apply in382_bad_closed y hy x
        simpa only [wild_joint_canonical_matching, ne_comm, min_comm, max_comm] using hxy
  6. Canonical types match and alpha values differ.

    Both triples contain the matching edge (1,2) and exclude 0, giving identical canonical type 2. For {1,2,11}, all permutations are bad so alpha = 1. For {1,2,3}, the good permutation is not bad, so bad permutations are a strict subset, giving alpha < 1.

    lines 65–107

        wild_joint_canonical_type 4 (by omega) {1,2,11} =
          wild_joint_canonical_type 4 (by omega) {1,2,3} ∧
        wild_joint_alpha 4 {1,2,11} = 1 ∧ wild_joint_alpha 4 {1,2,3} < 1 := by
      classical
      have hedge₁ : ∃ x ∈ ({1,2,11} : Finset (Fin 12)),
          ∃ y ∈ ({1,2,11} : Finset (Fin 12)), wild_joint_canonical_matching 4 x y :=
        ⟨1,by simp,2,by simp,by norm_num [wild_joint_canonical_matching,Fin.ext_iff]⟩
      have hedge₂ : ∃ x ∈ ({1,2,3} : Finset (Fin 12)),
          ∃ y ∈ ({1,2,3} : Finset (Fin 12)), wild_joint_canonical_matching 4 x y :=
        ⟨1,by simp,2,by simp,by norm_num [wild_joint_canonical_matching,Fin.ext_iff]⟩
      refine ⟨?_,?_,?_⟩
      · simp only [wild_joint_canonical_type, if_pos hedge₁, if_pos hedge₂]
        norm_num [Fin.ext_iff]
      · let D : Finset (Equiv.Perm (Fin 12)) := univ.filter fun σ =>
          wild_joint_first_block 4 σ = {1,2,11}
        have hpos : 0 < D.card := card_pos.mpr ⟨in382_bad_perm,
          mem_filter.mpr ⟨mem_univ _,in382_bad_first⟩⟩
        have heq : (univ.filter fun σ : Equiv.Perm (Fin 12) =>
            wild_joint_first_block 4 σ = {1,2,11} ∧
            wild_joint_bad_relation 4 (wild_joint_block 4 σ) (wild_joint_canonical_matching 4)) = D := by
          ext σ
          simp only [D,mem_filter,mem_univ,true_and]
          exact ⟨And.left,fun h => ⟨h,in382_bad_condition_implies_bad σ h⟩⟩
        unfold wild_joint_alpha
        rw [heq]
        change (D.card : ℝ) / (D.card : ℝ) = 1
        exact div_self (ne_of_gt (by exact_mod_cast hpos))
      · let D : Finset (Equiv.Perm (Fin 12)) := univ.filter fun σ =>
          wild_joint_first_block 4 σ = {1,2,3}
        let N : Finset (Equiv.Perm (Fin 12)) := univ.filter fun σ =>
          wild_joint_first_block 4 σ = {1,2,3} ∧
            wild_joint_bad_relation 4 (wild_joint_block 4 σ) (wild_joint_canonical_matching 4)
        have hD : in382_good_perm ∈ D := mem_filter.mpr ⟨mem_univ _,in382_good_first⟩
        have hN : in382_good_perm ∉ N := by
          intro h
          exact in382_good_connected (mem_filter.mp h).2.2
        have hsub : N ⊆ D := by intro σ h; exact mem_filter.mpr ⟨mem_univ _,(mem_filter.mp h).2.1⟩
        have hlt : N.card < D.card := card_lt_card (Finset.ssubset_iff_subset_ne.mpr ⟨hsub,by
          intro h; exact hN (h.symm ▸ hD)⟩)
        have hpos : (0 : ℝ) < D.card := by exact_mod_cast card_pos.mpr ⟨in382_good_perm,hD⟩
        change (N.card : ℝ) / (D.card : ℝ) < 1
        apply (div_lt_one hpos).mpr
        exact_mod_cast hlt

Written by a model (gemini-3.8-flash) from the Lean proof and the goal states Lean reported at each step. Lean verified the proof; this explanation is a reading aid and is not itself machine-checked.

Lean source view module on GitHub

open Finsetopen scoped BigOperatorsset_option maxHeartbeats 4000000set_option maxRecDepth 10000def in382_good_perm : Equiv.Perm (Fin 12) where  toFun := ![1,2,3,0,4,5,6,7,9,8,10,11]  invFun := ![3,0,1,2,4,5,6,7,9,8,10,11]  left_inv := by intro x; fin_cases x <;> decide  right_inv := by intro x; fin_cases x <;> decidedef in382_bad_perm : Equiv.Perm (Fin 12) where  toFun := ![1,2,11,0,3,4,5,6,7,8,9,10]  invFun := ![3,0,1,4,5,6,7,8,9,10,11,2]  left_inv := by intro x; fin_cases x <;> decide  right_inv := by intro x; fin_cases x <;> decidetheorem in382_good_first : wild_joint_first_block 4 in382_good_perm = {1,2,3} := by  classical  ext x  simp only [wild_joint_first_block, mem_filter, mem_univ, true_and]  fin_cases x <;> norm_num [wild_joint_first_block,wild_joint_block,in382_good_perm, Fin.ext_iff]theorem in382_bad_first : wild_joint_first_block 4 in382_bad_perm = {1,2,11} := by  classical  ext x  simp only [wild_joint_first_block, mem_filter, mem_univ, true_and]  fin_cases x <;> norm_num [wild_joint_first_block,wild_joint_block,in382_bad_perm, Fin.ext_iff]theorem in382_good_connected :    ¬wild_joint_bad_relation 4 (wild_joint_block 4 in382_good_perm)      (wild_joint_canonical_matching 4) := by  unfold wild_joint_bad_relation wild_joint_block wild_joint_canonical_matching  decide +kerneltheorem in382_bad_closed : ∀ x ∈ ({1,2,11} : Finset (Fin 12)),    ∀ y, wild_joint_canonical_matching 4 x y → y ∈ ({1,2,11} : Finset (Fin 12)) := by  intro x hx y hxy  simp only [mem_insert,mem_singleton] at hx  rcases hx with rfl|rfl|rfl <;> fin_cases y <;>    norm_num [wild_joint_canonical_matching, Fin.ext_iff] at *theorem in382_bad_condition_implies_bad (σ : Equiv.Perm (Fin 12))    (hσ : wild_joint_first_block 4 σ = {1,2,11}) :    wild_joint_bad_relation 4 (wild_joint_block 4 σ) (wild_joint_canonical_matching 4) := by  classical  refine ⟨{0},by simp,by decide,?_⟩  have hmem : ∀ x, wild_joint_block 4 σ x ∈ ({0} : Finset (Fin 4)) ↔      x ∈ ({1,2,11} : Finset (Fin 12)) := by    intro x    rw [← hσ]    simp only [mem_singleton, wild_joint_first_block, mem_filter, mem_univ, true_and]    change (⟨(σ.symm x).val / 3, _⟩ : Fin 4) = 0 ↔ (σ.symm x).val < 3    rw [Fin.ext_iff]    simp only [Fin.val_zero, Nat.div_eq_zero_iff]    norm_num []  intro x y hxy  rw [hmem, hmem]  constructor  · exact fun hx => in382_bad_closed x hx y hxy  · intro hy    apply in382_bad_closed y hy x    simpa only [wild_joint_canonical_matching, ne_comm, min_comm, max_comm] using hxytheorem wild_joint_even_canonical_orbit_obstruction :    wild_joint_canonical_type 4 (by omega) {1,2,11} =      wild_joint_canonical_type 4 (by omega) {1,2,3} ∧    wild_joint_alpha 4 {1,2,11} = 1 ∧ wild_joint_alpha 4 {1,2,3} < 1 := by  classical  have hedge₁ : ∃ x ∈ ({1,2,11} : Finset (Fin 12)),      ∃ y ∈ ({1,2,11} : Finset (Fin 12)), wild_joint_canonical_matching 4 x y :=    ⟨1,by simp,2,by simp,by norm_num [wild_joint_canonical_matching,Fin.ext_iff]⟩  have hedge₂ : ∃ x ∈ ({1,2,3} : Finset (Fin 12)),      ∃ y ∈ ({1,2,3} : Finset (Fin 12)), wild_joint_canonical_matching 4 x y :=    ⟨1,by simp,2,by simp,by norm_num [wild_joint_canonical_matching,Fin.ext_iff]⟩  refine ⟨?_,?_,?_⟩  · simp only [wild_joint_canonical_type, if_pos hedge₁, if_pos hedge₂]    norm_num [Fin.ext_iff]  · let D : Finset (Equiv.Perm (Fin 12)) := univ.filter fun σ =>      wild_joint_first_block 4 σ = {1,2,11}    have hpos : 0 < D.card := card_pos.mpr ⟨in382_bad_perm,      mem_filter.mpr ⟨mem_univ _,in382_bad_first⟩⟩    have heq : (univ.filter fun σ : Equiv.Perm (Fin 12) =>        wild_joint_first_block 4 σ = {1,2,11} ∧        wild_joint_bad_relation 4 (wild_joint_block 4 σ) (wild_joint_canonical_matching 4)) = D := by      ext σ      simp only [D,mem_filter,mem_univ,true_and]      exact ⟨And.left,fun h => ⟨h,in382_bad_condition_implies_bad σ h⟩⟩    unfold wild_joint_alpha    rw [heq]    change (D.card : ℝ) / (D.card : ℝ) = 1    exact div_self (ne_of_gt (by exact_mod_cast hpos))  · let D : Finset (Equiv.Perm (Fin 12)) := univ.filter fun σ =>      wild_joint_first_block 4 σ = {1,2,3}    let N : Finset (Equiv.Perm (Fin 12)) := univ.filter fun σ =>      wild_joint_first_block 4 σ = {1,2,3} ∧        wild_joint_bad_relation 4 (wild_joint_block 4 σ) (wild_joint_canonical_matching 4)    have hD : in382_good_perm ∈ D := mem_filter.mpr ⟨mem_univ _,in382_good_first⟩    have hN : in382_good_perm ∉ N := by      intro h      exact in382_good_connected (mem_filter.mp h).2.2    have hsub : N ⊆ D := by intro σ h; exact mem_filter.mpr ⟨mem_univ _,(mem_filter.mp h).2.1⟩    have hlt : N.card < D.card := card_lt_card (Finset.ssubset_iff_subset_ne.mpr ⟨hsub,by      intro h; exact hN (h.symm ▸ hD)⟩)    have hpos : (0 : ℝ) < D.card := by exact_mod_cast card_pos.mpr ⟨in382_good_perm,hD⟩    change (N.card : ℝ) / (D.card : ℝ) < 1    apply (div_lt_one hpos).mpr    exact_mod_cast hlt
See it in the graphDiscuss this resultChallenge it