Let $(\Omega, \mathcal F,P)$ be a probability space, $I$ a countable subset of $\mathbb R$ and $(\mathcal F_i)_{i\in I}$ a filtration.

Let $\tau$ be a random variable with values in $I\cup \{\infty\}$. Prove that $\tau$ is a stopping time if and only if $\forall t\in I, (\tau=t)\in \mathcal F_t$

I've found the implication $\tau$ stopping time $\implies \forall t\in I, (\tau=t)\in \mathcal F_t$ surprisingly tedious to prove. Give some $t\in I$, I considered two cases, whether $t$ can be approached from below by elements of $I$ or not.

  • If it can, $$\displaystyle (\tau = t)= (\tau \leq t) \cap \bigcap_{\substack{a\in I\\a<t}} (\tau \leq a)^c \in \mathcal F_t$$ (note that countability of $I$ comes into play).

  • It it cannot, there is some $\epsilon >0$ such that $(t-\epsilon,t)\cap I =\emptyset$. If $(-\infty,t)\cap I =\emptyset$, we're done. Otherwise, I consider $u=\sup \{u<t, u\in I\}$. Then I checked two cases again:

    1. If $u\in I$, $(\tau = t) = (\tau \leq t) \cap (\tau > u) \in \mathcal F_t$

    2. If $u\notin I$, $$\displaystyle (\tau = t)= (\tau \leq t) \cap \bigcap_{\substack{a\in I\\a<u}} (\tau \leq a)^c \in \mathcal F_t$$

This proof is quite dull and lengthy, do you know a more concise one ?


The following proof is shorter, but the idea of the proof is actually the same:

Since $I$ is countable there exists an enumeration $(s_n)_{n \in J}$, $J \subseteq \mathbb{N}$ , of the set

$$\{s \in I; s < t\}.$$


$$\{\tau=t\} = \{\tau \leq t\} \cap \left( \bigcup_{n \in J}\{\tau \leq s_n\} \right)^c \in \mathcal{F}_t,$$

and this proves the assertion.

| cite | improve this answer | |

In the second case $\{\tau=t\}=\{\tau\le t\}\setminus\{\tau\le t-\epsilon\}$.

| cite | improve this answer | |

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.