Mathematics Stack Exchange is a question and answer site for people studying math at any level and professionals in related fields. Join them; it only takes a minute:

Sign up
Here's how it works:
  1. Anybody can ask a question
  2. Anybody can answer
  3. The best answers are voted up and rise to the top

I'm taking a course about Markov chain, and here's a snippet from the lecture notes:

Let $(X_i, i \ge 0)$ be a time homogeneous Markov chain, let $V$ be the state space, let $\lambda$ be the initial distribution, i.e. $P\{X_0 = v\} = \lambda(v)$. Let $T$ be a stopping time. Let $p_{w,v}$ be the transition probability from state $w$ to $v$.

Then for each $v \in V$ and $i \ge 1$, we have

\begin{align} & P_\lambda(X_i = v, T \ge i) \\ = & \sum_{w \in V} P_\lambda(X_{i-1}=w, T \ge i-1)\times P_\lambda(T>i-1|X_{i-1} = w, T \ge i-1)\times p_{w,v} \end{align}

From the context, I can see there's no typo in the equations, but when I try to derive it my self, I can only get

\begin{align} & \sum_{w \in V} P_\lambda(X_{i-1}=w, T \ge i-1)\times P_\lambda(T>i-1|X_{i-1} = w, T \ge i-1)\times p_{w,v} \\ = & \sum_{w \in V} P_\lambda(T>i-1, X_{i-1} =w, T \ge i-1) \times p_{w,v} \\ = & \sum_{w \in V} P_\lambda(T \ge i, X_{i-1} =w) \times p_{w,v} \\ = & \sum_{w \in V} P_\lambda(T \ge i, X_{i-1} =w) \times P(X_1=v|X_0=w) \end{align}

But I can't go any further. My guess is that, since $T$ is a stopping time, then we can use strong Markov property to get

$$ P_\lambda(X_i =v | T \ge i, X_{i-1}=w) = P(X_1=v|X_0=w). $$

Is this correct?

BTW, the original equation comes from a stopping rule called the "filling rule".

share|cite|improve this question
up vote 1 down vote accepted

The $p_{w,v}$ is meant to be the probability of going from $w$ at time $i-1$ to $v$ at time $i$. Once the system decides not to stop at time $i-1$, where it goes next is independent of everything (by definition of a stopping time), so that $P(X_i=v|X_{i-1}=w) = P(X_i=v|X_{i-1}=w,T\ge i)=p_{w,v}$. This gives you what you need.

share|cite|improve this answer

Your Answer


By posting your answer, you agree to the privacy policy and terms of service.

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