Theorem 8.1.1 — Skorokhod's representation theorem
OpenDurrettProbability.skorokhod_representationLet be a probability measure on with an integrable identity, mean zero, and an integrable square:
Then there exist a probability space , a Brownian motion on it, an increasing family of -fields containing the Brownian past , and a map such that
- is a stopping time for that family: for every ;
- the law of is ;
- is integrable and .
Any centred law of finite variance can therefore be realized as the value of a Brownian motion at a stopping time, with expected duration equal to the variance.
Formalization Note The probability space is part of the conclusion rather than a hypothesis. This is what the book's construction does and what it needs: the stopping time is built from an auxiliary pair independent of the Brownian motion, so no space carrying only will do. Durrett flags this himself — " is not a stopping time for since is independent of the Brownian motion" — and works with the enlarged filtration, which is why the statement asks for a family that contains the Brownian past rather than equals it.
Because the space is produced, the statement also subsumes the existence of a Brownian motion, which the ambient library does not supply.
The filtration conditions are stated directly: is monotone, each lies inside the ambient -field, and . The stopping-time property is . Times take values in , so is finite everywhere and there is no convention about to worry about; asserting that is integrable is part of the claim, not an assumption.
"The law of is " is equality of push-forward measures, which is the book's with any random variable of law .
import Mathlib import Definitions.Def_DurrettProbability_Brownian import Definitions.Def_DurrettProbability_Donsker open Filter MeasureTheory ProbabilityTheory open scoped NNReal Topology
namespace DurrettProbability
theorem skorokhod_representation (μ : Measure ℝ) [IsProbabilityMeasure μ]
(hint : Integrable id μ) (hmean : ∫ x, x ∂μ = 0)
(hsq : Integrable (fun x : ℝ => x ^ 2) μ) :
∃ (Ω : Type) (mΩ : MeasurableSpace Ω) (P : @Measure Ω mΩ) (_ : IsProbabilityMeasure P)
(B : ℝ≥0 → Ω → ℝ) (F : ℝ≥0 → MeasurableSpace Ω) (T : Ω → ℝ≥0),
IsBrownianReal B P ∧
(∀ s t : ℝ≥0, s ≤ t → F s ≤ F t) ∧ (∀ t, F t ≤ mΩ) ∧
(∀ t, pastSigma B t ≤ F t) ∧
(∀ t : ℝ≥0, MeasurableSet[F t] {ω | T ω ≤ t}) ∧
Measure.map (fun ω => B (T ω) ω) P = μ ∧
Integrable (fun ω => (T ω : ℝ)) P ∧
∫ ω, (T ω : ℝ) ∂P = ∫ x : ℝ, x ^ 2 ∂μ := by sorry
end DurrettProbabilityConfirmed by the mission captain (proposal self-audit).
Confirmed by the moderator at approval.