Doob's maximal inequality
ProvedFamousTheorems.doob_maximal_inequalityDoob's maximal inequality. Let be a nonnegative submartingale on a finite measure space with respect to a filtration, let , and let . Then
In particular . So the running maximum of a submartingale is controlled by its final value alone. This inequality is the source of Doob's inequalities and is a basic tool in martingale theory and stochastic analysis.
Formalization note. Mathlib's MeasureTheory.maximal_ineq. The running maximum is (Finset.range (n + 1)).sup' of the values , with the nonemptiness proof Finset.nonempty_range_add_one. The left side is in ENNReal, and the right side is ENNReal.ofReal of the set integral of . ε is a nonnegative real (NNReal).
Preamble
import Mathlib
Formal statement
namespace FamousTheorems
open MeasureTheory
theorem doob_maximal_inequality {Ω : Type*} {m0 : MeasurableSpace Ω} {μ : Measure Ω} {𝒢 : Filtration ℕ m0} {f : ℕ → Ω → ℝ}
[IsFiniteMeasure μ] (hsub : Submartingale f 𝒢 μ) (hnonneg : 0 ≤ f) {ε : NNReal} (n : ℕ) :
(ε : ENNReal) * μ {ω | (ε : ℝ) ≤ (Finset.range (n + 1)).sup' Finset.nonempty_range_add_one fun k => f k ω} ≤
ENNReal.ofReal (∫ ω in {ω | (ε : ℝ) ≤ (Finset.range (n + 1)).sup' Finset.nonempty_range_add_one fun k => f k ω},
f n ω ∂μ) := by sorry
end FamousTheoremsSource
Listed in Mathlib's curated theorem manifests (docs/1000.yaml); formalized in Mathlib as `MeasureTheory.maximal_ineq`. Proof here reduces to that Mathlib result.