Prove2Me
Navigate
DiscoverCollectionsFormalpediaGalaxyBlogsUsersMomentumMy Missions+
Prove2Me
Log in
← Formalpedia

B.9, p. 35 — (19) evaluated at α∗p\alpha^*pα∗p with ≤ implies α~(p)≥α∗p\tilde\alpha(p)\ge\alpha^*pα~(p)≥α∗p

Proved
SampleDrivenOS.Linear.b9_reduction

by mikedeng1 · Oct 15, 2026 · Mathlib 0df444a (Lean v4.33.1)

optimal-stoppingp2o-batch-pfp2ap2o-gran-per-chapterp2o-plan-paperp2o-v1real-analysis

Let p∈(0,1)p\in(0,1)p∈(0,1) and let aaa be a root of (19), i.e. 0<a<10<a<10<a<1 and

ln⁡p+a(1−p)2p=∑i=1∞ln⁡(1−a+a[(i+1)pi−ipi+1])i(i+1).\ln p+\frac{a(1-p)^2}{p}=\sum_{i=1}^{\infty}\frac{\ln\bigl(1-a+a[(i+1)p^i-ip^{i+1}]\bigr)}{i(i+1)}.lnp+pa(1−p)2​=i=1∑∞​i(i+1)ln(1−a+a[(i+1)pi−ipi+1])​.

Let s>0s>0s>0 with sp<1sp<1sp<1. If

ln⁡p+s(1−p)2≤∑i=1∞ln⁡(1−sp+sp[(i+1)pi−ipi+1])i(i+1),\ln p+s(1-p)^2\le\sum_{i=1}^{\infty}\frac{\ln\bigl(1-sp+sp[(i+1)p^i-ip^{i+1}]\bigr)}{i(i+1)},lnp+s(1−p)2≤i=1∑∞​i(i+1)ln(1−sp+sp[(i+1)pi−ipi+1])​,

then sp≤asp\le asp≤a.

Writing a=f pa=f\,pa=fp, equation (19) becomes ln⁡p+f(1−p)2=∑iln⁡(1−fp+fp[(i+1)pi−ipi+1])/(i(i+1))\ln p+f(1-p)^2=\sum_i\ln(1-fp+fp[(i+1)p^i-ip^{i+1}])/(i(i+1))lnp+f(1−p)2=∑i​ln(1−fp+fp[(i+1)pi−ipi+1])/(i(i+1)), whose left side increases and whose right side decreases in fff; so the displayed inequality at f=sf=sf=s forces s≤a/ps\le a/ps≤a/p. With s=α∗s=\alpha^*s=α∗ this reduces Theorem 9 to the displayed inequality.

Formalization Note The paper writes f(p)=α~(p)/pf(p)=\tilde\alpha(p)/pf(p)=α~(p)/p and concludes f(p)≥α∗f(p)\ge\alpha^*f(p)≥α∗; the statement is phrased as sp≤asp\le asp≤a to avoid the division and holds for every root aaa of (19). The hypothesis sp<1sp<1sp<1 keeps every logarithm's argument positive (at least 1−sp1-sp1−sp); α∗p<1\alpha^*p<1α∗p<1 holds since α∗<1\alpha^*<1α∗<1. The series are Lean tsums over i : ℕ with iii shifted by one; both are summable in this range.

Preamble
import Mathlib
import Definitions.Def_SampleDrivenOS_Linear_Setting
Formal statement
namespace SampleDrivenOS.Linear

theorem b9_reduction (p : ℝ) (hp : p ∈ Set.Ioo (0 : ℝ) 1) (a : ℝ) (ha : SampleDrivenOS.Limit.IsAlphaTilde p a)
    (s : ℝ) (hs0 : 0 < s) (hsp : s * p < 1)
    (h : Real.log p + s * (1 - p) ^ 2 ≤
      ∑' i : ℕ, Real.log (1 - s * p + s * p * ((i + 2) * p ^ (i + 1) - (i + 1) * p ^ (i + 2))) /
        ((i + 1) * (i + 2))) :
    s * p ≤ a := by sorry

end SampleDrivenOS.Linear
Source
Correa, Cristi, Epstein & Soto, Sample-Driven Optimal Stopping: From the Secretary Problem to the i.i.d. Prophet Inequality, arXiv:2011.06516v2, p. 35, B.9 (Proof of Theorem 9), first paragraph and its two displays
Human review
  • Endorsed by Shuze Chen · Oct 15, 2026

    Confirmed by the moderator at approval.

  • Endorsed by mikedeng1 · Oct 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