Prove2Me
Navigate
DiscoverCollectionsFormalpediaGalaxyBlogsUsersMomentumMy Missions+
Prove2Me
Log in
← Formalpedia

Lemma 6.1.4 — likelihood-ratio monotonicity of the binomial filter update

Proved
MDPFinance.POMDPFinance.lemma_6_1_4

by Shuze Chen · Sep 27, 2026 · Mathlib 0df444a (Lean v4.33.1)

bayesian-inferencemarkov-decision-processpartially-observablestochastic-order

For the binomial market with unknown up-probability θ\thetaθ, this lemma records the two facts that drive every subsequent monotonicity result of the section. (a) A single down-move followed by an up-move produces a more favorable posterior than a single down-move alone: in terms of the explicit sufficient-statistic update, p^(⋅∣m,n+1)≤lrp^(⋅∣m+1,n+1)\hat p(\cdot\mid m,n+1) \le_{lr} \hat p(\cdot\mid m+1, n+1)p^​(⋅∣m,n+1)≤lr​p^​(⋅∣m+1,n+1), i.e. observing one more up-move (holding the trial count fixed) shifts the posterior up in the likelihood-ratio order. (b) The belief-only value function dkd_kdk​ (MDPFinance.POMDPFinance. dPow) is monotone in the likelihood-ratio order on reachable posteriors: ρ≤lrρ′\rho \le_{lr} \rho'ρ≤lr​ρ′ implies dk(ρ)≤dk(ρ′)d_k(\rho) \le d_k(\rho')dk​(ρ)≤dk​(ρ′) for every kkk.

Formalization Note. Part (a) is stated directly on the explicit posterior-density formula p^(θ∣m,n)∝θm(1−θ)n−mp0(θ)\hat p(\theta\mid m,n) \propto \theta^m(1-\theta)^{n-m}p_0(\theta)p^​(θ∣m,n)∝θm(1−θ)n−mp0​(θ) (binomialPosteriorDensity) rather than through the general filter update Φ\PhiΦ, since it is exactly this closed form the book's own proof inspects. Part (b) is stated for arbitrary densities p, p' in DQ0D_{Q_0}DQ0​​ satisfying the likelihood-ratio hypothesis, turned into measures via measureOfDensity (Lebesgue density on Θ=(0,1)\Theta=(0,1)Θ=(0,1)), matching the general statement's own universal quantification over ρ,ρ′∈DQ0\rho,\rho' \in D_{Q_0}ρ,ρ′∈DQ0​​.

Moderation note. Stated for the binomial market (IsBinomialMarket), a prior density p0p_0p0​, 0<d<1+i<u0<d<1+i<u0<d<1+i<u and 0<γ<10<\gamma<10<γ<1; part a) is the likelihood ratio order of the two Bayes updates Φ(ρ,dˉ)\Phi(\rho,\bar d)Φ(ρ,dˉ), Φ(ρ,uˉ)\Phi(\rho,\bar u)Φ(ρ,uˉ) of the filter (LRMeasure), part b) is for the recursion (6.4) dBin on normalized posteriors. The draft's part b) quantified over an arbitrary filter market and used unnormalized posterior densities, for which the claim is false.

Preamble
import Mathlib
import Definitions.Def_MDPFinance_POMDPFinance_Filter
import Definitions.Def_MDPFinance_POMDPFinance_Binomial
import Definitions.Def_MDPFinance_POMDPFinance_PowerLogValue

open MeasureTheory ProbabilityTheory
Formal statement
namespace MDPFinance.POMDPFinance

/-- Lemma 6.1.4 (Bäuerle–Rieder, p. 179, PDF 192), for the binomial market with unknown
up-probability `θ \in Θ = (0,1)` (`IsBinomialMarket`: constant hidden parameter, returns `\bar u`
w.p. `θ`, `\bar d` w.p. `1-θ`), prior density `p_0`, up/down factors `0 < d < 1+i < u`,
`\bar u := u/(1+i) - 1`, `\bar d := d/(1+i) - 1`, power utility with `0 < γ < 1` (the subsection's
standing case). For `ρ = \hat μ(\cdot\mid m,n) \in D_{Q_0}` it holds: a) `Φ(ρ,\bar d) \le_{lr}
Φ(ρ,\bar u)`, the likelihood ratio order (Definition B.3.5) of the two Bayes updates of the
filter `Φ` (which, by `FilterOp.hPhi` at the binomial market, are `\hat μ(\cdot\mid m,n+1)` and
`\hat μ(\cdot\mid m+1,n+1)`, the book's proof identity). b) `ρ \le_{lr} ρ' \Rightarrow d_k(\rho)
\le d_k(\rho')` for the binomial recursion (6.4) `dBin` (every `k`, matching the book's own
induction on `n`, restated with `k` = stages remaining; `ρ`, `ρ'` are posterior densities in
`D_{Q_0}`, turned into probability measures via `measureOfDensity`). -/
theorem lemma_6_1_4 (M : FilterMarket ℝ 1) (Fd : FilterOp M) (u dn i : ℝ) (hdn : 0 < dn)
    (hd : dn < 1 + i) (hi : 1 + i < u) (hM : IsBinomialMarket M u dn i) (γ : ℝ) (hγ0 : 0 < γ)
    (hγ1 : γ < 1) (p0 : ℝ → ℝ) (hp0 : IsPriorDensity p0) (m n : ℕ) (hmn : m ≤ n) :
    LRMeasure (volume.restrict (Set.Ioo (0 : ℝ) 1))
        (Fd.Phi (measureOfDensity (binomialPosteriorDensity p0 m n)) (dbar dn i))
        (Fd.Phi (measureOfDensity (binomialPosteriorDensity p0 m n)) (ubar u i)) ∧
      (∀ p p' : ℝ → ℝ, p ∈ DQ0 p0 → p' ∈ DQ0 p0 → LikelihoodRatioOrder p p' →
        ∀ k : ℕ, dBin M Fd u dn i γ k (measureOfDensity p) ≤
          dBin M Fd u dn i γ k (measureOfDensity p')) := by sorry

end MDPFinance.POMDPFinance
Source
Bäuerle and Rieder, Markov Decision Processes with Applications to Finance, Universitext, Springer 2011, DOI 10.1007/978-3-642-18324-9, p. 179, Lemma 6.1.4
Human review
  • Endorsed by Community (Bot) · Oct 1, 2026

    Confirmed by the moderator at approval.

  • Endorsed by Shuze Chen · Oct 1, 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