Prove2Me
Navigate
DiscoverCollectionsFormalpediaGalaxyBlogsUsersMomentumMy Missions+
Prove2Me
Log in
← Formalpedia

Theorem 2, p. 1215 — for i.i.d. Xᵢ ≥ 0, sup_n sup_X EX*ₙ/sup_{T*ₙ} E⁺X_t = sup_n sup_X EX*ₙ/sup_{T*ₙ} EX_t = 2

Proved
SamuelCahnProphet.IID.theorem_2

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

iidoptimal-stoppingp2o-batch-pfp2ap2o-gran-per-chapterp2o-plan-paperp2o-v1prophet-inequalitythreshold-rule

Let X1,…,XnX_1, \dots, X_nX1​,…,Xn​ be i.i.d. nonnegative random variables, Xn∗=max⁡(X1,…,Xn)X_n^* = \max(X_1, \dots, X_n)Xn∗​=max(X1​,…,Xn​), and let Tn∗T_n^*Tn∗​ be the class of threshold rules t(c)t(c)t(c), s(c)s(c)s(c) with c≥0c \ge 0c≥0. Theorem 2 of Samuel-Cahn (1984) states

sup⁡nsup⁡X‾[EXn∗sup⁡t∈Tn∗E+Xt]=sup⁡nsup⁡X‾[EXn∗sup⁡t∈Tn∗EXt]=2,\sup_n \sup_{\underline X} \left[\frac{EX_n^*}{\sup_{t \in T_n^*} E^+X_t}\right] = \sup_n \sup_{\underline X} \left[\frac{EX_n^*}{\sup_{t \in T_n^*} EX_t}\right] = 2,nsup​X​sup​[supt∈Tn∗​​E+Xt​EXn∗​​]=nsup​X​sup​[supt∈Tn∗​​EXt​EXn∗​​]=2,

the inner suprema ranging over all i.i.d. nonnegative X‾=(X1,…,Xn)\underline X = (X_1, \dots, X_n)X​=(X1​,…,Xn​). It is stated here as three facts:

  1. for every n≥1n \ge 1n≥1 and every law μ\muμ on [0,∞)[0, \infty)[0,∞), EXn∗≤2sup⁡t∈Tn∗E+XtEX_n^* \le 2 \sup_{t \in T_n^*} E^+X_tEXn∗​≤2supt∈Tn∗​​E+Xt​;
  2. for every n≥1n \ge 1n≥1 and every law, sup⁡t∈Tn∗E+Xt≤sup⁡t∈Tn∗EXt\sup_{t \in T_n^*} E^+X_t \le \sup_{t \in T_n^*} EX_tsupt∈Tn∗​​E+Xt​≤supt∈Tn∗​​EXt​;
  3. for every ε>0\varepsilon > 0ε>0 there are n≥1n \ge 1n≥1 and a law μ\muμ on [0,∞)[0, \infty)[0,∞) with (2−ε)sup⁡t∈Tn∗EXt<EXn∗(2 - \varepsilon) \sup_{t \in T_n^*} EX_t < EX_n^*(2−ε)supt∈Tn∗​​EXt​<EXn∗​.

Facts 1 and 2 bound both ratios by 222; fact 3 shows that the second ratio exceeds 2−ε2 - \varepsilon2−ε, and by fact 2 so does the first. So both suprema equal 222: the factor 222 of Theorem 1 cannot be improved for threshold rules, even for i.i.d. variables.

Formalization Note. The i.i.d. variables are the coordinates of Rn\mathbb R^nRn under the product measure μ⊗n\mu^{\otimes n}μ⊗n (the standard i.i.d. model; all quantities depend only on the joint law), so the supremum over X‾\underline XX​ is a supremum over probability laws μ\muμ with μ((−∞,0))=0\mu((-\infty, 0)) = 0μ((−∞,0))=0. Expectations are in [0,∞][0, \infty][0,∞]. The ratios are not written literally, because in [0,∞][0, \infty][0,∞] the quotients 0/00/00/0 and ∞/∞\infty/\infty∞/∞ take junk values; the three facts give the page's two equalities wherever the ratios are defined. The strict inequality in fact 3 excludes witnesses with EXn∗=0EX_n^* = 0EXn∗​=0 or sup⁡tEXt=∞\sup_t EX_t = \inftysupt​EXt​=∞. In fact 3 the sample size is written n+1n + 1n+1 with n∈Nn \in \mathbb Nn∈N, so that it is at least 111. Fact 2 holds for every measure and is stated without the probability hypothesis.

Preamble
import Mathlib
import Definitions.Def_SamuelCahnProphet_IID_Setting

open MeasureTheory Filter Topology
Formal statement
namespace SamuelCahnProphet.IID

theorem theorem_2 :
    (∀ (n : ℕ) [NeZero n] (μ : Measure ℝ) [IsProbabilityMeasure μ], μ (Set.Iio 0) = 0 →
        Emax μ n ≤ 2 * supEplus μ n) ∧
      (∀ (n : ℕ) [NeZero n] (μ : Measure ℝ), supEplus μ n ≤ supE μ n) ∧
      (∀ ε : ℝ, 0 < ε → ∃ (n : ℕ) (μ : Measure ℝ), IsProbabilityMeasure μ ∧ μ (Set.Iio 0) = 0 ∧
        ENNReal.ofReal (2 - ε) * supE μ (n + 1) < Emax μ (n + 1)) := by sorry

end SamuelCahnProphet.IID
Source
Samuel-Cahn, Comparison of threshold stop rules and maximum for independent nonnegative random variables, Ann. Probab. 12 (1984), p. 1215, Theorem 2
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