Kolmogorov's zero–one law specifies that a certain type of event, namely a tail event of independent σ-algebras, will either almost surely happen or almost surely not happen; that is, the probability of such an event occurring is zero or one.
Motivation
Motivation
LetA1={ω:n=1∑∞nξnconverges}be the set of sample points for which ∑n=1∞(ξ/n) converges (to a finite number) and consider the probability P(A1) of this set. It is far from clear, to begin with, what values this probability might have. However, it is a remarkable fact that we are able to say that the probability can have only two values, 0 or 1. This is a corollary of Kolmogorov's "zero-one law".
Tail σ-algebra T
Tail σ-algebra T
Let (Ω,F,P) be a probability space, and let ξ1,ξ2,... be a sequence of random variable. Let Fn∞=σ(ξn,ξn+1,...) be the σ-algebra generated by ξn,ξn+1,..., and writeT=n=1⋂∞Fn∞.Since the intersection of σ-algebras is again a σ-algebra, T is a σ-algebra. It is called a ---tail algebra---, because every event A∈T is independent of the values of ξ1,...,ξn for every finite number n, and is determined, so to speak, only by the behavior of the infinitely remote values of ξ1,ξ2,..... \\Since, for every k≥1,A1={n=1∑∞nξnconverges}={n=k∑∞nξnconverges}∈Fk∞we have A1∈⋂kFk∞≡T. In the same way, if ξ1,ξ2,.... is any sequence,A2={n=1∑∞ξnconverges}∈T
Theorem
Theorem
Theorem (Kolmogorov's Zero-One Law) Let ξ1,ξ2,.... be a sequence of independent random variables and let A∈T. The P(A) can only have one of the values zero and one.
The idea of the proof is to show that every tail event A is independent of itself and therefore P(A∩A)=P(A)⋅P(A), i.e., P(A)=P2(A) so that P(A)=0 or 1. If A∈F then A∈F1∞=σ{ξ1,ξ2,...}=σ(⋃nF1n), where F1n=σ{ξ1,ξ2,...,ξn}, and we can find sets An∈F1n,m≥1, such that P(A△An)→0,n→0. HenceP(An)→P(A),P(A∩An)→P(A)But if A∈T, the events An and A are independent for every n≥1. Hence it follows that P(A)=P2(A) and therefore P(A)=0 or 1. □