Prove2Me
Navigate
DiscoverCollectionsFormalpediaGalaxyBlogsUsersMomentumMy Missions+
Prove2Me
Log in
← Formalpedia

Claim 8.2, p. 50 — trimming moves each player's mixed strategy by at most 2/(k − 1) in L1

Proved
RubinsteinNash.WeakNash.claim_8_2

by mikedeng1 · Oct 9, 2026 · Mathlib 0df444a (Lean v4.33.1)

approximate-nash-equilibriump2o-batch-pfp2ap2o-gran-per-chapterp2o-plan-paperp2o-v1weak-nash

Let xxx be a mixed-strategy profile of a finite game, let ϵ>0\epsilon>0ϵ>0 and k>1k>1k>1, and let x^\hat xx^ be the trimmed profile of the proof of Lemma 8.1 (each ϵ\epsilonϵ-optimal player vvv drops the actions more than ϵk\epsilon kϵk below Umax⁡v(x−v)U^v_{\max}(x^{-v})Umaxv​(x−v) and renormalises; every other player keeps xvx^vxv). Then for every player vvv,

∑s∈Sv∣x^sv−xsv∣ ≤ 2k−1.\sum_{s\in S^v}\bigl|\hat x^v_s-x^v_s\bigr|\ \le\ \frac{2}{k-1}.s∈Sv∑​​x^sv​−xsv​​ ≤ k−12​.

So when kkk is large the trimmed profile is close to xxx in every coordinate, which is what keeps the expected payoffs of all players nearly unchanged.

Formalization Note. The page states the claim for the players of the bipartite polymatrix game; it uses no polymatrix structure, so it is stated for an arbitrary finite game. For a player that is not ϵ\epsilonϵ-optimal, x^v=xv\hat x^v=x^vx^v=xv and the left side is 000.

Preamble
import Mathlib
import Definitions.Def_agt_games
import Definitions.Def_DGPNash_NashMap_nashMap
import Definitions.Def_DGPNash_WellSupported_Equilibria
import Definitions.Def_RubinsteinNash_WeakNash_ApproxEquilibria
Formal statement
namespace RubinsteinNash.WeakNash

theorem claim_8_2 {ι : Type*} [Fintype ι] [DecidableEq ι]
    {S : ι → Type*} [∀ i, Fintype (S i)] [∀ i, DecidableEq (S i)]
    (u : ι → (∀ i, S i) → ℝ) (x : ∀ i, S i → ℝ) (hx : AGT.IsMixedProfile x)
    (ε k : ℝ) (hε : 0 < ε) (hk : 1 < k) :
    ∀ v : ι, ∑ s : S v, |weakTrim u x ε k v s - x v s| ≤ 2 / (k - 1) := by sorry

end RubinsteinNash.WeakNash
Source
Rubinstein, arXiv:1606.04550 (version dated August 26, 2016), Claim 8.2 (= [32, Claim 6]), §8.1, p. 50
Human review
  • Endorsed by Shuze Chen · Oct 9, 2026

    Confirmed by the moderator at approval.

  • Endorsed by mikedeng1 · Oct 9, 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