Kolmogorov's 0-1 law
ProvedFamousTheorems.kolmogorov_zero_one_lawKolmogorov's 0-1 law. Let be a sequence of independent sub-σ-algebras on a probability space . Every event in the tail σ-algebra
has probability or .
For instance, for independent random variables , the convergence of and the event are tail events and so are almost sure or almost impossible. The law is a basic tool of probability theory, used for example in the Borel–Cantelli lemmas and the strong law of large numbers.
Formalization note. Mathlib's ProbabilityTheory.measure_zero_or_one_of_measurableSet_limsup_atTop. The tail σ-algebra is Filter.limsup s Filter.atTop in the complete lattice of σ-algebras, which equals the intersection above. ProbabilityTheory.iIndep s μ is mutual independence and already forces μ to be a probability measure.
import Mathlib
namespace FamousTheorems
open scoped MeasureTheory
theorem kolmogorov_zero_one_law {Ω : Type*} {m0 : MeasurableSpace Ω} {μ : MeasureTheory.Measure Ω} (s : ℕ → MeasurableSpace Ω)
(h_le : ∀ n, s n ≤ m0) (h_indep : ProbabilityTheory.iIndep s μ) {t : Set Ω}
(ht_tail : MeasurableSet[Filter.limsup s Filter.atTop] t) : μ t = 0 ∨ μ t = 1 := by sorry
end FamousTheorems