Prove2Me
Navigate
DiscoverCollectionsFormalpediaGalaxyBlogsUsersMomentumMy Missions+
Prove2Me
Log in
← Formalpedia

Lemma 5.4 — Markov-chain recovery

Proved
SecondBestTrade.HalfEfficiency.markov_recovery

by mikedeng1 · Nov 8, 2026 · Mathlib 0df444a (Lean v4.33.1)

algorithmic-game-theorymechanism-designp2o-batch-stoc27ap2o-gran-per-chapterp2o-plan-paperp2o-v1

Fix a useful dual. For each buyer and seller aaa, let QaQ_aQa​ be its constructed generator and let Va,V~aV_a,\widetilde V_aVa​,Va​ be its true and virtual coordinate matrices. Then V~a=(I−Qa)Va\widetilde V_a=(I-Q_a)V_aVa​=(I−Qa​)Va​, I−QaI-Q_aI−Qa​ is invertible, and

Ra=(I−Qa)−1=∫0∞e−θeθQa dθ.R_a=(I-Q_a)^{-1}=\int_0^\infty e^{-\theta}e^{\theta Q_a}\,d\theta.Ra​=(I−Qa​)−1=∫0∞​e−θeθQa​dθ.

The matrices eθQae^{\theta Q_a}eθQa​ form a continuous stochastic semigroup and are the transition matrices of a finite-state continuous-time Markov chain from each initial support type. Every row is a probability distribution. Stopping independently at a rate-one exponential time gives row Ra(t,⋅)R_a(t,\cdot)Ra​(t,⋅) as the terminal law. For every incident coordinate eee,

E[V~a(t^,e)∣t]=∑sRa(t,s)V~a(s,e)=Va(t,e).\mathbb E[\widetilde V_a(\widehat t,e)\mid t]=\sum_s R_a(t,s)\widetilde V_a(s,e)=V_a(t,e).E[Va​(t,e)∣t]=s∑​Ra​(t,s)Va​(s,e)=Va​(t,e).

The resolvent entries and exponential-weighted virtual-coordinate expectations are integrable. The conclusion applies to every buyer and every seller, with the same generator acting on all coordinates of an agent.

This recovers entire multidimensional types and includes the chain and terminal-law interpretation as well as the algebraic identity.

Differences from the source:

  • encoding: finite agent, trade and support sets with product-prior expectations → Fin labels, an injective nonnegative-vector support embedding, and finite product-weighted sums; reason: bijections cover every finite source market, and summing an overwritten own coordinate multiplies by its total probability one
  • encoding: matrix integral, independently exponentially stopped chain, and vector recovery using (5.8) → entrywise integrable density integrals, finite-dimensional chain laws on a probability space, stochastic continuous semigroup, and coordinatewise recovery with zero extension outside incident trades; reason: rate-one stopping has density exp(-theta); finite coordinates give the same integral, law and vector identities; (5.8) is written explicitly and no recovery is assumed
Preamble
import Mathlib
import Definitions.Def_SecondBestTrade_HalfEfficiency_Recovery

open scoped BigOperators NNReal
open MeasureTheory
set_option autoImplicit false
noncomputable section
Formal statement
namespace SecondBestTrade.HalfEfficiency

/-- Lemma 5.4, printed p. 10 (PDF p. 11), with both sides and every coordinate. -/
theorem markov_recovery (N : Market) (z : Dual N) (hz : useful N z) :
    ∀ a : Agent N,
      (∀ t e, virtual N z a t e = ∑ s,
        (1 - generator N z a) t s * coord N a s e) ∧
      recoveryProperties (generator N z a) (coord N a) (virtual N z a) := by sorry

end SecondBestTrade.HalfEfficiency
Source
Dirk Bergemann, Yang Cai, Kangning Wang, Jinzhao Wu, Renzhe Xu, STOC 2027 submission #529 (October 2026), Lemma 5.4 — Markov-chain recovery, printed p. 10 (PDF p. 11)
Human review
  • Endorsed by Shuze Chen · Nov 8, 2026

    Confirmed by the moderator at approval.

  • Endorsed by mikedeng1 · Nov 8, 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