Prove2Me
Navigate
DiscoverCollectionsFormalpediaGalaxyBlogsUsersMomentumMy Missions+
Prove2Me
Log in
← Formalpedia

Proposition 3.1 — for ε = γ = 0 and ρ > 0 the worst-case conditional expected loss equals sup_{y∈𝒴} ℓ(y)

Proved
CondPortOT.SingUninf.proposition_3_1

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

conditional-expectationdistributionally-robust-optimizationoptimal-transportp2o-batch-pfp2ap2o-gran-per-chapterp2o-plan-paperp2o-v1portfolio-optimization

Let X\mathcal XX (covariates) and Y\mathcal YY (returns) be Polish spaces with Borel σ\sigmaσ-algebras, let (x^i,y^i)(\widehat x_i,\widehat y_i)(xi​,y​i​), i∈[N]i\in[N]i∈[N], N≥1N\ge1N≥1, be the data with empirical measure P^\widehat{\mathbb P}P, let x0∈Xx_0\in\mathcal Xx0​∈X, and let Bρ\mathbb B_\rhoBρ​ be the optimal transport ball of radius ρ\rhoρ around P^\widehat{\mathbb P}P for the ground cost DX+DY\mathbb D_{\mathcal X}+\mathbb D_{\mathcal Y}DX​+DY​. Suppose Assumption 2.4 holds with fiber radius γ=0\gamma=0γ=0 and that x0x_0x0​ is not isolated for DX\mathbb D_{\mathcal X}DX​. Then for every ρ>0\rho>0ρ>0 and every loss ℓ:Y→R\ell:\mathcal Y\to\mathbb Rℓ:Y→R,

sup⁡Q∈Bρ, Q(X=x0)>0EQ[ℓ(Y)∣X=x0]=sup⁡y∈Yℓ(y),\sup_{\mathbb Q\in\mathbb B_\rho,\ \mathbb Q(X=x_0)>0}\mathbb E_{\mathbb Q}[\ell(Y)\mid X=x_0]=\sup_{y\in\mathcal Y}\ell(y),Q∈Bρ​, Q(X=x0​)>0sup​EQ​[ℓ(Y)∣X=x0​]=y∈Ysup​ℓ(y),

with both sides in [−∞,+∞][-\infty,+\infty][−∞,+∞].

This is the case ε=γ=0\varepsilon=\gamma=0ε=γ=0 of the robustified conditional portfolio problem (4), with the mass constraint taken as the strict inequality Q(X=x0)>0\mathbb Q(X=x_0)>0Q(X=x0​)>0. The result is negative: the worst-case conditional loss does not depend on the data and equals the robust (support-function) value, so conditioning on a single covariate value with an optimal transport ball is uninformative.

Formalization Note ℓ\ellℓ stands for y↦ℓ(y,α,β)y\mapsto\ell(y,\alpha,\beta)y↦ℓ(y,α,β) at an arbitrary fixed decision (α,β)(\alpha,\beta)(α,β), so the theorem quantifies over all functions ℓ:Y→R\ell:\mathcal Y\to\mathbb Rℓ:Y→R; no measurability, boundedness or integrability is assumed, and the conditional expectation uses the extended integral. "X=x0X=x_0X=x0​" is the fiber {x:DX(x,x0)≤0}×Y\{x:\mathbb D_{\mathcal X}(x,x_0)\le0\}\times\mathcal Y{x:DX​(x,x0​)≤0}×Y, equal to {x0}×Y\{x_0\}\times\mathcal Y{x0​}×Y under Assumption 2.4(i). The hypothesis that x0x_0x0​ is not isolated (for every r>0r>0r>0 some x≠x0x\ne x_0x=x0​ has DX(x,x0)≤r\mathbb D_{\mathcal X}(x,x_0)\le rDX​(x,x0​)≤r) is an addition to the paper: it is Assumption 2.4(iii) at γ=0\gamma=0γ=0 when the boundary of {x0}\{x_0\}{x0​} is taken in an ambient Euclidean space, as the paper's remark on p. 11 suggests, and the proof uses it; with the boundary taken in X\mathcal XX itself, Assumption 2.4(iii) is vacuous at an isolated point, where the statement fails.


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_CondPortOT_SingUninf_Setting

open MeasureTheory
Formal statement
namespace CondPortOT.SingUninf

theorem proposition_3_1 {X Y : Type*} [TopologicalSpace X] [PolishSpace X] [MeasurableSpace X]
    [BorelSpace X] [TopologicalSpace Y] [PolishSpace Y] [MeasurableSpace Y] [BorelSpace Y]
    (DX : X → X → ℝ) (DY : Y → Y → ℝ) {N : ℕ} (hN : 0 < N) (xh : Fin N → X) (yh : Fin N → Y)
    (x0 : X) (hA : Assumption24 DX DY x0 0 xh) (hH1 : NotIsolated DX x0)
    (ρ : ℝ) (hρ : 0 < ρ) (L : Y → ℝ) :
    worstCasePos DX DY xh yh x0 0 ρ L = ⨆ y, (L y : EReal) := by sorry

end CondPortOT.SingUninf
Source
Nguyen et al., Robustifying conditional portfolio decisions via optimal transport, arXiv:2103.16451v3, p. 14, Proposition 3.1 (setting: §3.1, p. 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