AFTD Auto-Formalizing Theoretical Domains

publication_choice_binary_equilibrium_not_unique

theoremverifiederratum

Theorem 4.1 of arXiv:2511.13678 is false: not every binary-type Publication Choice Problem satisfying the standing assumptions, Assumption 1 and Assumption 2 has a unique pure-strategy equilibrium.

Statement

theorem publication_choice_binary_equilibrium_not_unique :
    ¬ ∀ (α β : ℝ) (θ μ : Fin 2 → ℝ) (c : Fin 2 → Fin 2 → ℝ),
        pcp_is_valid α β θ μ c → pcp_monotone_cost_ratio θ c → pcp_noncompetitive_venue c →
        ∀ (a a' : Fin 2 → Fin 2 → ℝ) (v v' : Fin 2 → ℝ),
          pcp_is_equilibrium α β θ μ c a v → pcp_is_equilibrium α β θ μ c a' v' → v = v'
Corrects
The Publication Choice Problem, AAAI 2026 (arXiv:2511.13678v2), Theorem 4.1 (a binary-type problem satisfying the standing assumptions and Assumptions 1-2 has a unique pure-strategy equilibrium): false; counterexample alpha = 1/2, beta = 2, two venues, types (1, 8), three equilibria
Verified
27 Oct 2026
Axioms
Classical.choiceQuot.soundpropext
Built on 5
  • pcp_is_equilibriumPure-strategy equilibrium of a Publication Choice Problem (Definition 1): every researcher type best responds to the venue impacts, and…
  • pcp_is_validStanding assumptions of a Publication Choice Problem: alpha in (0,1), beta > 1, positive type densities, types and costs, and costs…
  • pcp_monotone_cost_ratioAssumption 1 (Monotone Cost Ratio): for types theta_i < theta_i' and venues j < j', c_ij / c_i'j < c_ij' / c_i'j'.
  • pcp_noncompetitive_venueAssumption 2 (non-competitive venue): the least competitive venue has the same publication cost for every type.
  • publication_choice_three_equilibriaA binary-type Publication Choice Problem with two venues, alpha = 1/2, beta = 2, satisfying Assumptions 1 and 2, has three pure-strategy…

Read back from the Lean

The declaration asserts that equilibrium venue impacts are not always unique in 2-author, 2-venue publication choice models: specifically, it is false that for all real parameters α and β, all author type distributions θ, μ : Fin 2 → ℝ, and all cost matrices c : Fin 2 → Fin 2 → ℝ satisfying the conditions: 1. pcp_is_valid α β θ μ c (validity: 0 < α < 1, β > 1, positive measures μ, positive qualities θ, positive costs c, and strictly higher costs for venue 1 than venue 0 for each author type), 2. pcp_monotone_cost_ratio θ c (strictly monotone cost ratios across author types and venues), and 3. pcp_noncompetitive_venue c (venue 0 has identical cost across both author types), any two equilibria (a, v) and (a', v') for the same parameters must have identical venue impact vectors v = v'.

Written by a model that saw only the Lean, never the English above. If the two disagree, that is worth a challenge.

Lean source view module on GitHub

/-- Theorem 4.1 of arXiv:2511.13678 is false: not every binary-type Publication Choice Problem satisfying the standing assumptions, Assumption 1 (MCR) and Assumption 2 (non-competitive venue) has a unique pure-strategy equilibrium, even counting only the venue impacts. -/theorem publication_choice_binary_equilibrium_not_unique :    ¬ ∀ (α β : ℝ) (θ μ : Fin 2 → ℝ) (c : Fin 2 → Fin 2 → ℝ),        pcp_is_valid α β θ μ c → pcp_monotone_cost_ratio θ c → pcp_noncompetitive_venue c →        ∀ (a a' : Fin 2 → Fin 2 → ℝ) (v v' : Fin 2 → ℝ),          pcp_is_equilibrium α β θ μ c a v → pcp_is_equilibrium α β θ μ c a' v' → v = v' := by  intro h  obtain ⟨hv, hm, hn, a₁, a₂, _, v₁, v₂, _, q1, q2, _, _, h12, _⟩ := publication_choice_three_equilibria  have := h _ _ _ _ _ hv hm hn a₁ a₂ v₁ v₂ q1 q2  rw [this] at h12  exact lt_irrefl _ h12
See it in the graphDiscuss this resultChallenge it