Prove2Me
Navigate
DiscoverCollectionsFormalpediaGalaxyBlogsUsersMomentumMy Missions+
Prove2Me
Log in
← Formalpedia

Proposition 2, p. 7 — below the capacity threshold k̲ M enters [v_K, ∞) (Mode I); above it he also enters [v_L, v_T) (Mode II)

Proved
RetailerMarketplace.Entry.proposition_2

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

capacity-allocationmarketplacep2o-batch-pfp2ap2o-gran-per-chapterp2o-plan-paperp2o-v1product-entry

Let c>0c>0c>0 be the unit cost, f∈[0,1)f\in[0,1)f∈[0,1) the ad valorem fee, λ∈[12,1]\lambda\in[\tfrac12,1]λ∈[21​,1] the share of consumers MMM can steer (Assumption 1), and k∈[0,1]k\in[0,1]k∈[0,1] MMM's capacity. Let μ\muμ be the distribution of product valuations: a probability measure on R\mathbb RR with μ((−∞,c))=0\mu((-\infty,c))=0μ((−∞,c))=0, absolutely continuous with respect to Lebesgue measure, with ∫∣v∣ dμ<∞\int|v|\,d\mu<\infty∫∣v∣dμ<∞, and charging every nonempty open interval of [c,∞)[c,\infty)[c,∞); G(v)=μ((−∞,v])G(v)=\mu((-\infty,v])G(v)=μ((−∞,v]). Third-party sellers enter the products with v≥vT=c/(1−f)v\ge v_T=c/(1-f)v≥vT​=c/(1−f), and MMM chooses an allocation: any measurable set AAA of products with μ(A)≤k\mu(A)\le kμ(A)≤k, maximizing his total profit (SBM profits on AAA below vTv_TvT​, SBMT profits on AAA above vTv_TvT​, fee revenue on the marketplace products outside AAA).

Then there is a threshold k‾∈[0,1]\underline k\in[0,1]k​∈[0,1], namely k‾=1−G(v∗)\underline k=1-G(v^*)k​=1−G(v∗), such that

  1. Mode I (k≤k‾k\le\underline kk≤k​): an optimal allocation exists and is unique up to μ\muμ-null sets; it is ∅\emptyset∅ if k=0k=0k=0 and otherwise
A=[vK,∞)with1−G(vK)=k, vK≥vT,A=[v_K,\infty)\quad\text{with}\quad 1-G(v_K)=k,\ v_K\ge v_T,A=[vK​,∞)with1−G(vK​)=k, vK​≥vT​,

so products in [vT,vK)[v_T,v_K)[vT​,vK​) are supplied under SBT and products in [vK,∞)[v_K,\infty)[vK​,∞) under SBMT; 2. Mode II (k>k‾k>\underline kk>k​): there are c≤vL≤vT≤vHc\le v_L\le v_T\le v_Hc≤vL​≤vT​≤vH​ with

πMSBM(vL)=ΔπMSBMT−SBT(vH),(1−G(vH))+(G(vT)−G(vL))=k,\pi_M^{SBM}(v_L)=\Delta\pi_M^{SBMT-SBT}(v_H),\qquad (1-G(v_H))+(G(v_T)-G(v_L))=k,πMSBM​(vL​)=ΔπMSBMT−SBT​(vH​),(1−G(vH​))+(G(vT​)−G(vL​))=k,

such that A=[vL,vT)∪[vH,∞)A=[v_L,v_T)\cup[v_H,\infty)A=[vL​,vT​)∪[vH​,∞) is an optimal allocation and every optimal allocation equals it up to a μ\muμ-null set; so products in [vL,vT)[v_L,v_T)[vL​,vT​) are supplied under SBM, in [vT,vH)[v_T,v_H)[vT​,vH​) under SBT, and in [vH,∞)[v_H,\infty)[vH​,∞) under SBMT.

This is the paper's characterization of the two marketplace modes: with little capacity the monopolist competes only in the most valuable marketplace products; with more capacity he also monopolizes products no third party supplies.

Formalization Note Products below vTv_TvT​ are not marketplace products (TTT stays out), so an allocation avoiding them leaves them unsupplied. f<1f<1f<1 is assumed because vT=c/(1−f)v_T=c/(1-f)vT​=c/(1−f) is finite only then; the foreclosed case f=1f=1f=1 is not covered. c>0c>0c>0, integrability of vvv (so that profits are finite) and full support on [c,∞)[c,\infty)[c,∞) (the reading of vK=G−1(1−k)v_K=G^{-1}(1-k)vK​=G−1(1−k) in (3)) are added hypotheses. At k=0k=0k=0, vK=G−1(1)=∞v_K=G^{-1}(1)=\inftyvK​=G−1(1)=∞ and the optimal allocation is empty. At k=1k=1k=1 the Mode II thresholds are vL=cv_L=cvL​=c, vH=vTv_H=v_TvH​=vT​, so the bounds are closed. At f=0f=0f=0, k‾=1\underline k=1k​=1 and Mode II is vacuous, as in the paper.


Retired 2026-10-27: a Lean-vs-paper fidelity audit found this statement weaker than, or narrower than, the cited result. A corrected statement will be resubmitted.

Preamble
import Mathlib
import Definitions.Def_RetailerMarketplace_Entry_Model

open MeasureTheory
Formal statement
namespace RetailerMarketplace.Entry

/-- Proposition 2: there is a capacity threshold `k̲ ∈ [0, 1]` (namely `1 - G(v*)`) such that M's
optimal capacity allocations (unique up to `μ`-null sets) are
1. (Mode I, `k ≤ k̲`) the top products `[v_K, ∞)` with `1 - G(v_K) = k` and `v_K ≥ v_T`, so that
   `[v_T, v_K)` is SBT and `[v_K, ∞)` is SBMT (no product at all when `k = 0`);
2. (Mode II, `k > k̲`) `[v_L, v_T) ∪ [v_H, ∞)` with `c ≤ v_L ≤ v_T ≤ v_H`, where `v_L, v_H` equalize
   `π_M^SBM(v_L) = Δπ_M^{SBMT-SBT}(v_H)` and use all capacity, so that `[v_L, v_T)` is SBM,
   `[v_T, v_H)` is SBT and `[v_H, ∞)` is SBMT. -/
theorem proposition_2 (μ : Measure ℝ) [IsProbabilityMeasure μ] (c f lam k : ℝ) (hc : 0 < c)
    (hf : f ∈ Set.Ico (0 : ℝ) 1) (hlam : lam ∈ Set.Icc (1 / 2 : ℝ) 1)
    (hk : k ∈ Set.Icc (0 : ℝ) 1) (hsupp : μ (Set.Iio c) = 0) (hac : μ ≪ volume)
    (hint : Integrable (fun v : ℝ => v) μ)
    (hfull : ∀ a b : ℝ, c ≤ a → a < b → 0 < μ (Set.Ioo a b)) :
    ∃ k₀ ∈ Set.Icc (0 : ℝ) 1, k₀ = kbar μ c f lam ∧
      (k ≤ k₀ →
        ∃ U : Set ℝ,
          ((k = 0 ∧ U = ∅) ∨
            ∃ vK : ℝ, vT c f ≤ vK ∧ 1 - G μ vK = k ∧ U = Set.Ici vK) ∧
          IsOptimalAlloc μ c f lam k U ∧
          ∀ A : Set ℝ, IsOptimalAlloc μ c f lam k A → A =ᵐ[μ] U) ∧
      (k₀ < k →
        ∃ vL vH : ℝ, c ≤ vL ∧ vL ≤ vT c f ∧ vT c f ≤ vH ∧
          piSBM c vL = dPi c f lam vH ∧
          (1 - G μ vH) + (G μ (vT c f) - G μ vL) = k ∧
          IsOptimalAlloc μ c f lam k (Set.Ico vL (vT c f) ∪ Set.Ici vH) ∧
          ∀ A : Set ℝ, IsOptimalAlloc μ c f lam k A →
            A =ᵐ[μ] (Set.Ico vL (vT c f) ∪ Set.Ici vH : Set ℝ)) := by sorry

end RetailerMarketplace.Entry
Source
Hervas-Drane & Shelegia, Retailer-Led Marketplaces, Management Science (2025), Articles in Advance, doi:10.1287/mnsc.2023.00315, p. 7 (PDF p. 9), Proposition 2; display (3); proof pp. 13–15 (PDF pp. 15–17)
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