AFTD Auto-Formalizing Theoretical Domains

cycle_power_dyadic_full_prune_fractional_counterexample

theoremverified

Seventeen genuine tight radius-64 dyadic rooted profiles with positive integer multiplicities totaling1192 have expected size7, exact weighted dyadic pair counts1192 at every rank, and ordinary pair counts at least1192 at every distance1 through63, but every profile has a missing offset in its compressed outer quarters. Thus no positive full-pruned-class density follows from the independent single-class expectation constraints alone. This is an auxiliary fractional obstruction and is not an actual integer class model or a counterexample to OP-205.

Statement

theorem cycle_power_dyadic_full_prune_fractional_counterexample :
    ∃ (b : Fin 17 → ℕ → Bool) (w : Fin 17 → ℕ),
      (∑ i, w i) = 1192 ∧
      (∀ i, 0 < w i) ∧
      (∀ i, b i 0 = true ∧ b i 63 = true) ∧
      (∀ i, cycle_power_dyadic_class_weight 6 (b i) =
        (Finset.range 64).sum (fun a => (b i a).toNat)) ∧
      (∑ i, w i * (Finset.range 64).sum (fun a => (b i a).toNat)) = 8344 ∧
      (∀ j : Fin 6, (∑ i, w i * (Finset.range (2^(j : ℕ))).sum
        (fun a => (b i a && b i (a+64-2^(j : ℕ))).toNat)) = 1192) ∧
      (∀ d : Fin 63, 1192 ≤ ∑ i, w i * (Finset.range (64-((d : ℕ)+1))).sum
        (fun a => (b i a && b i (a+((d : ℕ)+1))).toNat)) ∧
      (∀ i, ∃ a : Fin 32, cycle_power_dyadic_b_lift 4 (b i) (a : ℕ) = false)
Source
Uniform OP-205 / RT-1 route audit following arXiv:2610.15635. Experiments7281/7291 recover a rational feasible mixture for the new full-prune density objective. The statement refutes only a sufficient linear-relaxation route and claims no graph covering value or research advance.
Verified
28 Oct 2026
Axioms
Classical.choiceQuot.soundpropext
Built on 2

The proof, explained

The theorem is proved by explicitly constructing a fractional counterexample: 17 dyadic rooted profiles of length 64 together with positive integer multiplicities summing to 1192. Each profile is defined as a 64-bit integer mask and its weight is given by an explicit integer vector. After instantiating the existential witnesses, the proof delegates the verification of all eight algebraic and structural properties to Lean's `decide` tactic, which checks them by direct finite computation.

  1. We construct explicit candidates for the 17 profiles b and their multiplicity weights w, reducing the existential goal to verifying the eight required constraints.

    The 17 profiles b are defined using 17 concrete 64-bit integer bitmasks, with bit a of profile i obtained by testing the a-th bit of the corresponding mask. The multiplicity weights w are given by an explicit list of 17 positive integers (596, 298, ..., 7). Providing these concrete witnesses with refine reduces the existential theorem to the conjunction of the eight stated properties.

    lines 18–21

      let masks : Fin 17 → ℕ := ![9223372036854775809, 13835058055282163715, 17293822569102704655, 18374686479671623935, 18442240474082185215, 18446462598732873727, 18446532961034551295, 18446480174812684287, 18410574545644814079, 18446726491188231167, 18444281149942726655, 18446714404645895679, 18446475740526460927, 18442203211885707263, 18442239243696734207, 18437487261123149823, 18446739786778541823]
      let b : Fin 17 → ℕ → Bool := fun i a => (masks i).testBit a
      let w : Fin 17 → ℕ := ![596, 298, 149, 69, 1, 12, 1, 11, 1, 8, 1, 8, 11, 9, 6, 4, 7]
      refine ⟨b, w, ?_⟩
  2. All eight properties hold for the constructed profiles and weights, completing the proof. automation

    Because all quantifiers run over finite index types (Fin 17, Fin 6, Fin 63, and Fin 32) and all summations are bounded arithmetic operations over explicit integers and bit tests, the entire conjunction is a decidable proposition. The decide tactic verifies all eight conditions by finite computation: the multiplicities sum to 1192 and are strictly positive; every profile has bits 0 and 63 set; the dyadic class weight matches the Hamming weight for each profile; the weighted sum of Hamming weights is 8344 (giving average size 7); dyadic boundary pair counts equal 1192 at each rank 0 through 5; pair counts at all distances 1 through 63 are at least 1192; and every profile has a zero bit in its 32-bit outer-quarter compression.

    line 22

      decide

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 Finsetset_option maxHeartbeats 2000000set_option maxRecDepth 10000theorem cycle_power_dyadic_full_prune_fractional_counterexample :    ∃ (b : Fin 17 → ℕ → Bool) (w : Fin 17 → ℕ),      (∑ i, w i) = 1192 ∧      (∀ i, 0 < w i) ∧      (∀ i, b i 0 = true ∧ b i 63 = true) ∧      (∀ i, cycle_power_dyadic_class_weight 6 (b i) =        (Finset.range 64).sum (fun a => (b i a).toNat)) ∧      (∑ i, w i * (Finset.range 64).sum (fun a => (b i a).toNat)) = 8344 ∧      (∀ j : Fin 6, (∑ i, w i * (Finset.range (2^(j : ℕ))).sum        (fun a => (b i a && b i (a+64-2^(j : ℕ))).toNat)) = 1192) ∧      (∀ d : Fin 63, 1192 ≤ ∑ i, w i * (Finset.range (64-((d : ℕ)+1))).sum        (fun a => (b i a && b i (a+((d : ℕ)+1))).toNat)) ∧      (∀ i, ∃ a : Fin 32, cycle_power_dyadic_b_lift 4 (b i) (a : ℕ) = false) := by  let masks : Fin 17 → ℕ := ![9223372036854775809, 13835058055282163715, 17293822569102704655, 18374686479671623935, 18442240474082185215, 18446462598732873727, 18446532961034551295, 18446480174812684287, 18410574545644814079, 18446726491188231167, 18444281149942726655, 18446714404645895679, 18446475740526460927, 18442203211885707263, 18442239243696734207, 18437487261123149823, 18446739786778541823]  let b : Fin 17 → ℕ → Bool := fun i a => (masks i).testBit a  let w : Fin 17 → ℕ := ![596, 298, 149, 69, 1, 12, 1, 11, 1, 8, 1, 8, 11, 9, 6, 4, 7]  refine ⟨b, w, ?_⟩  decide
See it in the graphDiscuss this resultChallenge it