Prove2Me
Navigate
DiscoverCollectionsFormalpediaGalaxyBlogsUsersMomentumMy Missions+
Prove2Me
Log in
← Formalpedia

Section 2, Eq. (16) — the loss of Rˉ\bar RRˉ tends to 0 for any two populations with finite means

Proved
RobbinsSeqDesign.ForcedSampling.forced_sampling_loss_tendsto_zero

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

bandit-algorithmsp2o-batch-pfp2bp2o-gran-per-chapterp2o-plan-paperp2o-v1sequential-design

Let AAA and BBB be any two populations whose expectations α\alphaα and β\betaβ exist (are finite), and let

1=a1<a2<⋯ ,2=b1<b2<⋯1 = a_1 < a_2 < \cdots, \qquad 2 = b_1 < b_2 < \cdots1=a1​<a2​<⋯,2=b1​<b2​<⋯

be disjoint sequences of positive integers of density 000. Then Robbins' rule Rˉ\bar RRˉ (forced draws from AAA at the aaa's and from BBB at the bbb's, otherwise from the population with the larger running mean, ties to BBB) satisfies

lim⁡n→∞Ln(A,B,Rˉ)=max⁡(α,β)−lim⁡n→∞E ⁣(Snn)=0,\lim_{n\to\infty} L_n(A, B, \bar R) = \max(\alpha,\beta) - \lim_{n\to\infty} E\!\left(\frac{S_n}{n}\right) = 0,n→∞lim​Ln​(A,B,Rˉ)=max(α,β)−n→∞lim​E(nSn​​)=0,

where Ln(A,B,R)=max⁡(α,β)−E(Sn/n)L_n(A,B,R) = \max(\alpha,\beta) - E(S_n/n)Ln​(A,B,R)=max(α,β)−E(Sn​/n) is the loss of display (9).

This answers affirmatively Robbins' question (12) — whether some rule RRR has lim⁡nLn(A,B,R)=0\lim_n L_n(A,B,R) = 0limn​Ln​(A,B,R)=0 for every A,BA, BA,B — not only for coins but for any two populations with finite means, and it does so with one explicit rule, uniformly in the choice of the density-zero schedule. In later terminology: a forced-exploration greedy policy has regret o(n)o(n)o(n) on every two-armed bandit with integrable rewards.

Formalization Note. "For any A,BA, BA,B such that α,β\alpha, \betaα,β exist" is: for every two-armed StochasticBandit with both reward laws integrable; no other restriction on the class of populations is imposed, and α=β\alpha=\betaα=β is allowed. The statement is for the specific rule Rˉ\bar RRˉ and every admissible schedule, not merely the existence of some rule. The limit is along n→∞n \to \inftyn→∞ in N\mathbb NN.

Preamble
import Mathlib
import Definitions.Def_RobbinsSeqDesign_ForcedSampling_Rule
import Definitions.Def_RobbinsSeqDesign_ForcedSampling_Loss
Formal statement
namespace RobbinsSeqDesign.ForcedSampling

open MeasureTheory ProbabilityTheory Filter Topology BanditAlgorithm

/-- Robbins (1952), Section 2, Eq. (16), p. 532: for any two populations A, B such that
`α`, `β` exist (are finite) and any pair of sequences (13), the loss (9) of the rule `R̄`
tends to zero: `lim_{n→∞} L_n(A, B, R̄) = 0`. -/
theorem forced_sampling_loss_tendsto_zero (ν : StochasticBandit 2)
    (hint : ∀ i, Integrable id (ν.P i)) (σ : ForcedSchedule) :
    Tendsto (fun n : ℕ => loss ν (rule σ) n) atTop (𝓝 0) := by sorry

end RobbinsSeqDesign.ForcedSampling
Source
Robbins, Some aspects of the sequential design of experiments, Bull. Amer. Math. Soc. 58 (1952), p. 532, Section 2, Eq. (16) (answering (12))
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