5
$\begingroup$

Let $p \in [0 , \frac{1}{2}] $ and $\eta_{i}$ be i.i.d random variables and $P(\eta_{i}=1)=p$ and $P(\eta_{i}=-1)=1-p$ and $\mathcal F_{n}=\sigma(\eta_{1},\cdots,\eta_{n})$ and $X_{n}=\sum_{i=1}^{n}\eta_{i}$ and $Y_{n}=X_{T(-7) \wedge n} $ . show that

1) $(X_{n})_{n\in \mathbb N}$ and $(Y_{n})_{n\in \mathbb N}$ are supermartingale or matigale or submartingale.

2) $(Y_{n})_{n\in \mathbb N}$ is almost surely convergence to $-7$.

thanks for any help.

$\endgroup$
14
  • 1
    $\begingroup$ Wait, you cannot prove that X is a supermartingale? $\endgroup$
    – Did
    Commented Nov 25, 2013 at 17:23
  • $\begingroup$ @Did.i edit the question I dont know which one is true. $\endgroup$ Commented Nov 25, 2013 at 17:32
  • $\begingroup$ @Did.can you help me to solve this problem? $\endgroup$ Commented Nov 25, 2013 at 17:41
  • $\begingroup$ $E[X_{n+1}|\mathcal{F}_n]=E[\eta_{n+1}|\mathcal{F}_n]+E[X_n|\mathcal{F}_n]$. What can you figure out about each term? I bet that helps you with the first part of 1. Now what does T mean to you? $\endgroup$
    – htd
    Commented Nov 25, 2013 at 18:35
  • $\begingroup$ @Henrik..subscript in $Y_{n\in \mathbb N}$ means stopping time. $\endgroup$ Commented Nov 26, 2013 at 5:25

1 Answer 1

6
$\begingroup$

OP has solved question 1 himself, see the comments. Regarding question 2:

Let $T(-7)=\inf\{n\geq 1 : X_n\in (-\infty,-7] \}$. Let us first make a quick argument that $P(T(-7)<\infty)=1$ which maybe you know.

By the law of large numbers $X_n/n \to p-(1-p) < 0$ a.s. if $p\in[0,1/2)$ so $X_n \to - \infty$ a.s. giving $P(T(-7)<\infty)=1$. For $p=1/2$ it is a (i hope) well known fact that $P(X_n \text{ hits m before -7})=-7/(-7-m) \to 0$ and by monotone convergence $P(T(-7)<\infty)=1-\lim_{m} P(X_n \text{ hits m before -7})=1$.

Then $\lim_{n} Y_{n} = \lim _{n} X_{T(-7)\wedge n} =X_{T(-7)}=-7$ a.s per properties of hitting times.

$\endgroup$

You must log in to answer this question.

Not the answer you're looking for? Browse other questions tagged .