Theorem 3.A.2 — the absolute-deviation characterization of the convex order
ProvedStochasticOrders.Convex.convex_order_abs_iffconvex-orderprobabilitystochastic-orders
Let on and on be integrable random variables with . Then
This is a fully self-contained alternative to Theorem 3.A.1's tail-integral form, needing no external equation references: holds exactly when 's mean absolute deviation from every point dominates 's. (The function is called the potential of 's law, so this restates as a pointwise comparison of the two potentials.)
Formalization Note |X ω - a| is Mathlib's absolute value on ℝ; no separate "potential"
definition is introduced since the theorem's content is fully captured by the raw inequality.
Preamble
import Mathlib import Definitions.Def_StochasticOrders_Convex_ConvexOrder
Formal statement
namespace StochasticOrders.Convex
open MeasureTheory
/-- Theorem 3.A.2 (Shaked & Shanthikumar, *Stochastic Orders*, Springer 2007, p. 111): let `X`
and `Y` be two random variables with `E[X] = E[Y]`. Then `X ≤cx Y` iff `E|X - a| ≤ E|Y - a|` for
all `a ∈ ℝ`. -/
theorem convex_order_abs_iff {Ω Ω' : Type*} [MeasurableSpace Ω] [MeasurableSpace Ω']
(μ : Measure Ω) (ν : Measure Ω') [IsProbabilityMeasure μ] [IsProbabilityMeasure ν]
(X : Ω → ℝ) (Y : Ω' → ℝ) (hX : Integrable X μ) (hY : Integrable Y ν)
(hmean : ∫ ω, X ω ∂μ = ∫ ω, Y ω ∂ν) :
ConvexOrder μ ν X Y ↔ ∀ a : ℝ, ∫ ω, |X ω - a| ∂μ ≤ ∫ ω, |Y ω - a| ∂ν := by sorry
end StochasticOrders.Convex
Source
Shaked & Shanthikumar, Stochastic Orders, Springer 2007, p. 111, Theorem 3.A.2
Human review
Confirmed by the mission captain (proposal self-audit).
Confirmed by the mission captain (proposal self-audit).
Confirmed by the moderator at approval.