Let $(X_k)$ be a sequence of independent random variables, all defined on the probability space $(\Omega,\mathcal{A},P)$. Fix a positive integer $n$ and define $$ C_n=\sigma(X_1,X_2,\cdots,X_n),\quad D_n=\sigma(X_{n+1},X_{n+2},\cdots). $$

Here is my question:

How to show by definition that $C_n$ and $D_n$ are independent?

By the definition of independence of $X_k$, for any finite set of positive integers $I$, the $\sigma$-algebras $(\sigma(X_i))_{i\in I}$ are independent. Eventually, one needs to show that

$$ P(A\cap B)=P(A)P(B) $$ for all $A\in C_n$ and $B\in D_n$. How shall I go on?

  • 1
    $\begingroup$ It's a lot of work to type out so I'll just leave a comment, but a theorem in Billingsley's book is a slight generalization of this with essentially the same proof. I believe in appears in chapter four. If you're looking for just a hint, use the $\pi - \lambda$ lemma. $\endgroup$ – RandomWalker Aug 30 '16 at 21:58
  • 1
    $\begingroup$ Indeed, without knowing what you know and do not know, this is next to impossible to answer. $\endgroup$ – Did Aug 30 '16 at 22:41

The proof of Kolmogrov's zero-one law in Jacod and Protter's Probability Essentials (Theorem 10.6) takes this statement for granted and it is sort of "cheating".

The proof of Theorem 22.3 (the zero-one law) in the third edition of Billingsley's Probability and Measure gives details about how to prove this statement, which eventually boils down to the Dynkin's $\pi$-$\lambda$ Theorem in the book (Theorem 3.2 page 42).


Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.