Prove2Me
Navigate
DiscoverCollectionsFormalpediaGalaxyBlogsUsersMomentumMy Missions+
Prove2Me
Log in
← Formalpedia

Theorem 8(a), p. 803 (if, corrected) — E((X/t)ξ∣X>t)→(1−ξ/α)−1E((X/t)^\xi\mid X>t)\to(1-\xi/\alpha)^{-1}E((X/t)ξ∣X>t)→(1−ξ/α)−1 with a finite moment gives F∈Dr(Γα)F\in D_r(\Gamma_\alpha)F∈Dr​(Γα​)

Proved
BalkemaDeHaan.Moments.theorem_8a_if

by mikedeng1 · Oct 6, 2026 · Mathlib 0df444a (Lean v4.33.1)

extreme-value-theorymomentsp2o-batch-pfp2bp2o-gran-per-chapterp2o-plan-paperp2o-v1regular-variationresidual-life

Let XXX be a real random variable with distribution function FFF such that F(x)<1F(x) < 1F(x)<1 for all real xxx, and let 0<ξ<α0 < \xi < \alpha0<ξ<α. Suppose that ∫0∞yξ dF(y)\int_0^\infty y^\xi\, dF(y)∫0∞​yξdF(y) is finite and

lim⁡t→∞E((Xt)ξ ∣ X>t)=(1−ξα)−1.\lim_{t\to\infty} E\Big(\Big(\frac{X}{t}\Big)^{\xi}\,\Big|\, X > t\Big) = \Big(1 - \frac{\xi}{\alpha}\Big)^{-1}.t→∞lim​E((tX​)ξ​X>t)=(1−αξ​)−1.

Then lim⁡t→∞P{X/t≤x∣X>t}=Γα(x−1)\lim_{t\to\infty} P\{X/t \le x \mid X > t\} = \Gamma_\alpha(x-1)limt→∞​P{X/t≤x∣X>t}=Γα​(x−1) for every x>0x > 0x>0.

This is the sufficiency half of the moment characterization: convergence of a single conditional moment of order ξ\xiξ to the value (1−ξ/α)−1(1 - \xi/\alpha)^{-1}(1−ξ/α)−1 already forces the scaled residual life times to converge to the Pareto law with index α\alphaα.

Formalization Note The paper's sufficiency condition is only that ∫0∞yξdF(y)\int_0^\infty y^\xi dF(y)∫0∞​yξdF(y) be finite and that the limit ccc exist and be finite. As printed this is false: for the Pareto tail 1−F(x)=x−β1 - F(x) = x^{-\beta}1−F(x)=x−β (x≥1x \ge 1x≥1) with β>ξ\beta > \xiβ>ξ, β≠α\beta \ne \alphaβ=α, the moment is finite and E((X/t)ξ∣X>t)=(1−ξ/β)−1E((X/t)^\xi \mid X > t) = (1-\xi/\beta)^{-1}E((X/t)ξ∣X>t)=(1−ξ/β)−1 for all t≥1t \ge 1t≥1, yet the limit law is Γβ(x−1)≠Γα(x−1)\Gamma_\beta(x-1) \ne \Gamma_\alpha(x-1)Γβ​(x−1)=Γα​(x−1). The statement here requires the limit to be (1−ξ/α)−1(1 - \xi/\alpha)^{-1}(1−ξ/α)−1, the value the paper itself identifies in "Then c=(1−ξ/α)−1c = (1-\xi/\alpha)^{-1}c=(1−ξ/α)−1".

Preamble
import Mathlib
import Definitions.Def_BalkemaDeHaan_ParetoBounds_GammaLaw
import Definitions.Def_BalkemaDeHaan_Moments_ResidualLife

open MeasureTheory Filter Topology
Formal statement
namespace BalkemaDeHaan.Moments

/-- Theorem 8(a), p. 803, the "if" half, **corrected**: if `F(x) < 1` for all `x`, `0 < ξ < α`,
`∫_0^∞ y^ξ dF(y)` is finite and `E((X/t)^ξ | X > t) → (1 − ξ/α)^{−1}` as `t → ∞`, then
`P{X/t ≤ x | X > t} → Γ_α(x − 1)` for all `x > 0`. The page asks only that the limit `c` exist and be
finite; that is false (a Pareto BalkemaDeHaan.LimitTypes.tail `x^{−β}`, `β > ξ`, `β ≠ α`, gives `c = (1 − ξ/β)^{−1}` and the
limit law `Γ_β(x − 1)`), so the limit value `(1 − ξ/α)^{−1}` is part of the hypothesis. -/
theorem theorem_8a_if (μ : Measure ℝ) [IsProbabilityMeasure μ]
    (hD₀ : ∀ x : ℝ, 0 < μ (Set.Ioi x))
    (α ξ : ℝ) (hξ : 0 < ξ) (hξα : ξ < α)
    (hmom : IntegrableOn (fun y : ℝ => y ^ ξ) (Set.Ioi 0) μ)
    (hc : Tendsto (fun t : ℝ => condMoment μ ξ t) atTop (𝓝 (1 - ξ / α)⁻¹)) :
    ∀ x : ℝ, 0 < x →
      Tendsto (fun t : ℝ => scaledResidualCDF μ t x) atTop (𝓝 (BalkemaDeHaan.ParetoBounds.GammaLaw α (x - 1))) := by sorry

end BalkemaDeHaan.Moments
Source
Balkema, de Haan, Residual Life Time at Great Age, Ann. Probab. 2 (1974), p. 803 (PDF p. 12), Theorem 8(a), "if" direction (corrected: c = (1 − ξ/α)^{−1} assumed)
Human review
  • Endorsed by Shuze Chen · Oct 6, 2026

    Confirmed by the moderator at approval.

  • Endorsed by mikedeng1 · Oct 6, 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