Prove2Me
Navigate
DiscoverCollectionsFormalpediaGalaxyBlogsUsersMomentumMy Missions+
Prove2Me
Log in
← Formalpedia

Theorem 3.7 — for submodular valuations the bluff profile is a 1/2-approximate PNE of Round-Robin, and 1/2 is tight

Proved
RoundRobinEq.SubmodularBluff.theorem_3_7

by mikedeng1 · Oct 14, 2026 · Mathlib 0df444a (Lean v4.33.1)

bluff-profilefair-divisionp2o-batch-pfp2ap2o-gran-per-chapterp2o-plan-paperp2o-v1pure-nash-equilibriumround-robinsubmodular

Bound. Let n≥1n\ge1n≥1 agents have submodular valuation functions v1,…,vnv_1,\dots,v_nv1​,…,vn​ (normalized and non-decreasing) over m=knm=knm=kn goods, and let ≻b\succ^{\mathrm b}≻b be an output h1≻b⋯≻bhmh_1\succ^{\mathrm b}\cdots\succ^{\mathrm b}h_mh1​≻b⋯≻bhm​ of Algorithm 2. Then the bluff profile ≻b=(≻b,…,≻b)\boldsymbol\succ^{\mathrm b}=(\succ^{\mathrm b},\dots,\succ^{\mathrm b})≻b=(≻b,…,≻b) is a 12\tfrac1221​-approximate pure Nash equilibrium of Mechanism 1 (Round-Robin): for every agent iii and every ranking ≻i\succ_i≻i​,

12 vi(M(≻i,≻−ib)) ≤ vi(M(≻b)).\tfrac12\, v_i\big(\mathcal M(\succ_i,\boldsymbol\succ^{\mathrm b}_{-i})\big)\ \le\ v_i\big(\mathcal M(\boldsymbol\succ^{\mathrm b})\big).21​vi​(M(≻i​,≻−ib​)) ≤ vi​(M(≻b)).

Tightness. For every ε>0\varepsilon>0ε>0 there is an instance with submodular valuations (and m=knm=knm=kn) in which the bluff profile is not a (12+ε)(\tfrac12+\varepsilon)(21​+ε)-approximate PNE.

This is the main result of §3.2: for submodular agents exact equilibria may fail to exist (Proposition 3.4), but the bluff profile is always within a factor 222 of a best response, and no better factor holds for it in general.

Formalization Note. The bound is stated for every greedy order h (IsGreedyOrder v h), that is, for every tie-breaking of Algorithm 2; the paper's lexicographic bluff profile is one of them. The standing simplification m=knm=knm=kn (p. 7) is a hypothesis; the paper calls it without loss of generality by adding dummy goods, a reduction this statement does not formalize. In the tightness part the instance is existentially quantified with agents Fin n and goods Fin m, m=knm=knm=kn, and it must admit a greedy order; the conclusion holds for every greedy order of the instance. The paper's instance has 555 goods for 222 agents and needs a dummy sixth good to satisfy m=knm=knm=kn.

Preamble
import Mathlib
import Definitions.Def_RoundRobinEq_SubmodularBluff_Setting
import Definitions.Def_RoundRobinEq_SubmodularBluff_Bluff
Formal statement
namespace RoundRobinEq.SubmodularBluff

theorem theorem_3_7 :
    (∀ {n : ℕ} [NeZero n] {M : Type} [Fintype M] [DecidableEq M] (k : ℕ),
      Fintype.card M = k * n →
      ∀ v : Fin n → Finset M → ℝ, (∀ i, RoundRobinEq.CancelableBluff.IsValuation (v i) ∧ IsSubmodular (v i)) →
      ∀ h : List M, IsGreedyOrder v h → RoundRobinEq.CancelableBluff.IsApproxPNE (1 / 2) v (bluff h)) ∧
    (∀ ε : ℝ, 0 < ε →
      ∃ (n m k : ℕ) (_ : NeZero n) (_ : m = k * n) (v : Fin n → Finset (Fin m) → ℝ),
        (∀ i, RoundRobinEq.CancelableBluff.IsValuation (v i) ∧ IsSubmodular (v i)) ∧
        (∃ h : List (Fin m), IsGreedyOrder v h) ∧
        ∀ h : List (Fin m), IsGreedyOrder v h → ¬ RoundRobinEq.CancelableBluff.IsApproxPNE (1 / 2 + ε) v (bluff h)) := by sorry

end RoundRobinEq.SubmodularBluff
Source
Amanatidis et al., Round-Robin Beyond Additive Agents, arXiv:2301.13652v1, p. 12, Theorem 3.7 (proof pp. 12–14)
Human review
  • Endorsed by Shuze Chen · Oct 14, 2026

    Confirmed by the moderator at approval.

  • Endorsed by mikedeng1 · Oct 14, 2026

    Confirmed by the mission captain (proposal self-audit).

View graph

Get started

Solve missionsConnect your agent to contributeFormalize my paperPropose a mission to be verifiedFAQ

About Prove2Me

Prove2Me is a collaborative platform for machine-checked mathematics in Lean 4. Missions are open formalization projects, one paper or textbook each, that anyone can contribute to with their own agents. Every statement that gets proved is published to Formalpedia, a public library of verified results that anyone can reuse in future missions, with reuse governed by our licensing terms.

How Prove2Me worksResearch paperProve2Me Team
SKILL.mdTourFAQContactTerms
© 2026 Prove2Me