Documentation

Atlas.TheoryOfProbability.code.OptionalStoppingAlt

theorem uniformIntegrable_stopped_second_term {Ω : Type u_1} {m0 : MeasurableSpace Ω} {μ : MeasureTheory.Measure Ω} {𝒢 : MeasureTheory.Filtration m0} {f : Ω} {N : Ω} (hf : MeasureTheory.Submartingale f 𝒢 μ) (hUI : MeasureTheory.UniformIntegrable f 1 μ) (hN : MeasureTheory.IsStoppingTime 𝒢 fun (ω : Ω) => (N ω)) :
MeasureTheory.UniformIntegrable (fun (n : ) (ω : Ω) => f (N ω) ω * {ω : Ω | N ω n}.indicator 1 ω) 1 μ

Helper lemma: if f is a uniformly integrable submartingale adapted to 𝒢 and N is a stopping time, then the sequence n ↦ f_N · 1_{N ≤ n} (the contribution to the stopped process from indices that have already been stopped) is uniformly integrable in .

theorem optional_stopping_alt_ui_preserved {Ω : Type u_1} {m0 : MeasurableSpace Ω} {μ : MeasureTheory.Measure Ω} {𝒢 : MeasureTheory.Filtration m0} {f : Ω} {N : Ω} (hf : MeasureTheory.Submartingale f 𝒢 μ) (hUI : MeasureTheory.UniformIntegrable f 1 μ) (hN : MeasureTheory.IsStoppingTime 𝒢 fun (ω : Ω) => (N ω)) :
MeasureTheory.UniformIntegrable (fun (n : ) (ω : Ω) => f (min (N ω) n) ω) 1 μ

General optional stopping preservation of uniform integrability: if f is a uniformly integrable submartingale adapted to 𝒢 and N is a stopping time, then the stopped process n ↦ f_{min(N, n)} is also uniformly integrable. This is the key ingredient for the general optional stopping theorem (Lecture 29).