Suppose $U_1,...,U_n$ are independent random variable with $\mathbb{E}[U_i]=0$.

Define $Z_k:=\sum_{i=1}^k U_i$. Set $T:=\inf \lbrace k \in N \mid |Z_k|>2\alpha \rbrace$.

Clearly $\lbrace T=k \rbrace$ is independent of $(Z_n-Z_k)=\sum_{i=k+1}^n U_i$.

I do not get why:

$$\Pr(|Z_n|>\alpha)\geq \Pr(|Z_n-Z_T|\leq \alpha, T\leq n)$$


By the definition of $T$ we have $Z_T>2\alpha$ whenever $T<\infty$. If, in addition, $|Z_n-Z_T|\leq \alpha$ we get

$$|Z_n|-\alpha=\alpha + |Z_n|-2\alpha > \alpha + |Z_n|-|Z_T| \geq \alpha - |Z_n-Z_T|\geq 0.$$

  • $\begingroup$ thank you very much, it is very clear. However I don't get the last inequality. Why should it be: $|Z_n|-|Z_T|\geq -|Z_N-Z_T|$ ? I see it is similar to the reverse triangular inequality but in that I have $\mid |Z_n|-|Z_T|\mid$ $\endgroup$ – mastro Jul 26 '15 at 8:41
  • $\begingroup$ It is just the triangle inequality. Work on it for awhile... :) $\endgroup$ – user940 Jul 26 '15 at 14:10

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.