I have got a Markov Chain as below
$\begin{bmatrix}q & p & 0 & 0 &0\\0 & p &q & 0 &0\\0&p&0&q&0\\q&0&0&0&p\\0&0&0&0&1\end{bmatrix}$
I am asked to classify the class and if they are transient. The answer from my lecturer is {0,1,2,3} is a open class so that its transient; and {4} is closed so that its persistent. What i am struggling is, for the class {0,1,2,3}, for example taking state 1, it can get back to state 1 in the next step, otherwise it can go to state 2, then it either back to state 1 or go to state 3, then it either go to state 4 which the matrix will terminate, or it can go to state 0 which can go back to state 1. So i can conclude that if we start in state 1, there are possibilities that it will go back to state 1 so that {0,1,2,3} which include state 1 shouldn't be transient.
Are there any mistake in my argument or i have mess up the definition of a transient state?
Many thanks
A state is transient if there's a nonzero proability that, starting there, that state is never visited again. State 1 in your chain is transient, because from state 1 you can go (with probability $q$) to state 2, then (with probability $q$) to state 3, then (with probability $p$) to state 4, after which you will never return to state 1; so if you start at 1 the probability of never returning to 1 is at least $pq^2$ which is bigger than 0.
I think you may be thinking that "transient" means "nonzero probability of returning", but that's not right: "recurrent" means "zero probability of not returning" and "transient" means "nonzero probability of not returning" or "not-1 probability of returning".