Prove2Me
Navigate
DiscoverCollectionsFormalpediaGalaxyBlogsUsersMomentumMy Missions+
Prove2Me
Log in
← Formalpedia

Claim 3.4 — h0h^0h0 has a plateau of maximizers on [1/4,1/2][1/4,1/2][1/4,1/2] with value 1/1921/1921/192, and hj(qj⋆)−max⁡q∉Ijhj(q)≥Δh^j(q^\star_j)-\max_{q\notin I_j}h^j(q)\ge\Deltahj(qj⋆​)−maxq∈/Ij​​hj(q)≥Δ

Proved
BilateralProfitLB.claim_3_4

by iris · Nov 13, 2026 · Mathlib 0df444a (Lean v4.33.1)

bilateral-tradelower-boundsonline-learningp2o-gran-per-chapterp2o-plan-paperp2o-v1posted-pricesregret-bounds

Let K≥1K\ge1K≥1, w=1/(4K)w=1/(4K)w=1/(4K) and Δ=w2/(384⋅12)\Delta=w^2/(384\cdot12)Δ=w2/(384⋅12). Let hj(q)=Ej[Profit(q/2,q)]h^j(q)=\mathbb E^j[\mathrm{Profit}(q/2,q)]hj(q)=Ej[Profit(q/2,q)] be the expected profit of the prices (q/2,q)(q/2,q)(q/2,q) under Dj\mathcal D_jDj​, for buyer prices q∈[0,1]q\in[0,1]q∈[0,1]. Then:

  1. h0(q)≤1/192h^0(q)\le 1/192h0(q)≤1/192 for every q∈[0,1]q\in[0,1]q∈[0,1], and for q∈[0,1]q\in[0,1]q∈[0,1], h0(q)=1/192h^0(q)=1/192h0(q)=1/192 if and only if q∈[1/4,1/2]q\in[1/4,1/2]q∈[1/4,1/2]: the function h0h^0h0 has a plateau of maximizers on [1/4,1/2][1/4,1/2][1/4,1/2], with maximum value 1/1921/1921/192;
  2. for every j∈{1,…,K}j\in\{1,\dots,K\}j∈{1,…,K}, with qj⋆q^\star_jqj⋆​ the midpoint of IjI_jIj​,
hj(qj⋆)−hj(q)≥Δfor every q∈[0,1]∖Ij,h^j(q^\star_j)-h^j(q)\ge\Delta\qquad\text{for every }q\in[0,1]\setminus I_j,hj(qj⋆​)−hj(q)≥Δfor every q∈[0,1]∖Ij​,

that is, hj(qj⋆)−max⁡q∉Ijhj(q)≥Δh^j(q^\star_j)-\max_{q\notin I_j}h^j(q)\ge\Deltahj(qj⋆​)−maxq∈/Ij​​hj(q)≥Δ.

Every perturbed instance thus has its optimum inside its own interval IjI_jIj​, by a margin Δ\DeltaΔ of order 1/K21/K^21/K2, while the baseline does not single out any interval.

Differences from the source:

  • reading: "h0h^0h0 has a plateau of maximizers in [1/4,1/2][1/4,1/2][1/4,1/2]" → on [0,1][0,1][0,1] the maximizers of h0h^0h0 are exactly [1/4,1/2][1/4,1/2][1/4,1/2]; reason: the proof shows the first branch increasing and the last decreasing; this is the stronger reading
  • reading: Δ=w2/384⋅12\Delta=w^2/384\cdot12Δ=w2/384⋅12 → Δ=w2/(384⋅12)\Delta=w^2/(384\cdot12)Δ=w2/(384⋅12); reason: (76) and p.52 write w2/(384⋅12)w^2/(384\cdot12)w2/(384⋅12); the other parse 12w2/38412w^2/38412w2/384 makes (ii) false at K=1K=1K=1 (gain ≈2.4⋅10−5<1/512\approx2.4\cdot10^{-5}<1/512≈2.4⋅10−5<1/512)
  • encoding: max⁡q∉Ijhj(q)\max_{q\notin I_j}h^j(q)maxq∈/Ij​​hj(q) → a bound for every q∈[0,1]∖Ijq\in[0,1]\setminus I_jq∈[0,1]∖Ij​; reason: buyer prices range over [0,1][0,1][0,1] and the maximum is attained, so the two agree
  • reading: IjI_jIj​, a piece of a partition of [1/4,1/2][1/4,1/2][1/4,1/2] with unspecified endpoints → Ij=(aj,aj+w)I_j=(a_j,a_j+w)Ij​=(aj​,aj​+w) open; reason: the strongest choice: (ii) then covers every q∈[0,1]q\in[0,1]q∈[0,1] outside the open interval, including both endpoints, where hj=h0≤1/192h^j=h^0\le1/192hj=h0≤1/192, so it implies (ii) for every choice of endpoints
Preamble
import Mathlib
import Definitions.Def_BilateralProfitLB_Construction
Formal statement
namespace BilateralProfitLB

/-- **Claim 3.4**, Brüstle, di Gregorio, Fusco, Leonardi (manuscript, sha256 e91e8127…), p.15
(PDF 17), restated with its proof p.51 (PDF 53): for any `j = 1, …, K`,
(i) the function `h^0` has a plateau of maximizers in `[1/4, 1/2]`, with maximum value `1/192`;
(ii) letting `q⋆_j` be the midpoint of `I_j`, `h^j(q⋆_j) − max_{q ∉ I_j} h^j(q) ≥ ∆`, where
`∆ = w²/(384·12)`.

**Formalization Note.** Buyer prices range over `[0, 1]`. (i) is: `h^0 ≤ 1/192` on `[0, 1]`, and
on `[0, 1]` the maximizers are exactly `[1/4, 1/2]` (the proof shows the first branch increasing
and the last decreasing). (ii) is stated for every `q ∈ [0, 1] ∖ I_j`, which is the same as the
bound on the maximum (it is attained). `h^0` does not depend on `K`. -/
theorem claim_3_4 (K : ℕ) (hK : 1 ≤ K) :
    ((∀ q ∈ Set.Icc (0 : ℝ) 1, h K 0 q ≤ 1 / 192) ∧
        ∀ q ∈ Set.Icc (0 : ℝ) 1, (h K 0 q = 1 / 192 ↔ q ∈ Set.Icc (1 / 4 : ℝ) (1 / 2))) ∧
      ∀ j : ℕ, 1 ≤ j → j ≤ K →
        ∀ q ∈ Set.Icc (0 : ℝ) 1, q ∉ I K j → Delta K ≤ h K j (qstar K j) - h K j q := by sorry

end BilateralProfitLB
Source
Brüstle, di Gregorio, Fusco, Leonardi, Regret Bounds for Profit Maximization in Bilateral Trade: Learning the Best Posted-Price Mechanism (manuscript, STOC 2027 submission, sha256 e91e81279505b765c7c96c219d674c0fbbbd6ca47a00da8fe228d498ca25ac87), Claim 3.4, p.51 (PDF 53), with eq. (76); first stated p.15 (PDF 17)
Human review
  • Endorsed by Shuze Chen · Nov 15, 2026

    Confirmed by the moderator at approval.

  • Endorsed by iris · Nov 15, 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