Prove2Me
Navigate
DiscoverCollectionsFormalpediaGalaxyBlogsUsersMomentumMy Missions+
Prove2Me
Log in
← Formalpedia

Theorem 3 — for (f₁ + ϵ)n ≤ k ≤ (1 − fₘ − ϵ)n, every non-adaptive policy has regret at least M√n

Proved
MultiSecretary.NonAdaptive.sqrt_regret_lower_bound

by mikedeng1 · Oct 5, 2026 · Mathlib 0df444a (Lean v4.33.1)

lower-boundnon-adaptive-policyp2o-batch-pfp2bp2o-gran-per-chapterp2o-plan-paperp2o-v1regret

Let 0<am<am−1<⋯<a10<a_m<a_{m-1}<\dots<a_10<am​<am−1​<⋯<a1​ and ϵ>0\epsilon>0ϵ>0. There is a constant M=M(ϵ,m,a1,…,am)>0M=M(\epsilon,m,a_1,\dots,a_m)>0M=M(ϵ,m,a1​,…,am​)>0 such that for all masses fj>0f_j>0fj​>0 summing to one with ϵ=12min⁡{fm,…,f1}\epsilon=\tfrac12\min\{f_m,\dots,f_1\}ϵ=21​min{fm​,…,f1​} and all (n,k)(n,k)(n,k) with (f1+ϵ)n≤k≤(1−fm−ϵ)n(f_1+\epsilon)n\le k\le(1-f_m-\epsilon)n(f1​+ϵ)n≤k≤(1−fm​−ϵ)n,

Mn≤Voff∗(n,k)−Vna∗(n,k).M\sqrt n\le V^*_{\mathrm{off}}(n,k)-V^*_{\mathrm{na}}(n,k).Mn​≤Voff∗​(n,k)−Vna∗​(n,k).

When the budget is a fraction of nnn bounded away from f1f_1f1​ and from 1−fm1-f_m1−fm​, even the best non-adaptive policy loses order n\sqrt nn​ against the offline optimum, while adaptive policies (the Budget-Ratio policy of the same paper) achieve regret bounded uniformly in nnn and kkk.

Formalization Note Lean index j∈{0,…,m−1}j\in\{0,\dots,m-1\}j∈{0,…,m−1} is the paper's index j+1j+1j+1, so a 0 is the largest value a1a_1a1​, f 0 is f1f_1f1​ and f (Fin.rev 0) is fmf_mfm​. The constant is chosen after ϵ\epsilonϵ, mmm, aaa and before fff, nnn, kkk, and it is required to be strictly positive (with M=0M=0M=0 the claim would reduce to Vna∗≤Voff∗V^*_{\mathrm{na}}\le V^*_{\mathrm{off}}Vna∗​≤Voff∗​). The statement holds for all nnn in the range, as printed, with no threshold on nnn. The range is empty unless f1+fm+2ϵ≤1f_1+f_m+2\epsilon\le1f1​+fm​+2ϵ≤1, so the statement has content only for m≥3m\ge3m≥3.

Preamble
import Mathlib
import Definitions.Def_MultiSecretary_NonAdaptive_Model
import Definitions.Def_MultiSecretary_NonAdaptive_Policy
Formal statement
namespace MultiSecretary.NonAdaptive

open Finset

/-- Theorem 3 (p. 25): for `ϵ = ½ min f_j` and `(f_1 + ϵ) n ≤ k ≤ (1 - f_m - ϵ) n` there is a constant
`M ≡ M(ϵ, m, a_1, …, a_m) > 0` with `M√n ≤ V*_off(n, k) - V*_na(n, k)`. -/
theorem sqrt_regret_lower_bound (ε : ℝ) (hε : 0 < ε) (m : ℕ) [NeZero m] (a : Fin m → ℝ)
    (ha : IsValues a) :
    ∃ M : ℝ, 0 < M ∧ ∀ f : Fin m → ℝ, IsMasses f → eps f = ε → ∀ n k : ℕ,
      (f 0 + ε) * n ≤ k → (k : ℝ) ≤ (1 - f (Fin.rev 0) - ε) * n →
      M * Real.sqrt n ≤ Voff a f n k - VstarNa a f n k := by sorry

end MultiSecretary.NonAdaptive
Source
Arlotto, Gurvich, Uniformly Bounded Regret in the Multi-Secretary Problem, arXiv:1710.07719v2, Theorem 3, p. 25 (= Theorem 1, second display, p. 5)
Human review
  • Endorsed by Shuze Chen · Oct 5, 2026

    Confirmed by the moderator at approval.

  • Endorsed by mikedeng1 · Oct 5, 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