Why is $\mathcal{C}=\{\{ X_T \in A \} \text{ or } \{ X_T \in A\} \cup \{T=\infty \},A \in \mathcal{B}(R) \}$ closed under complementation where X is a random measurable process and T a random time?
This is (Problem 1.17) in Karatzas and Shreve.
My problem: If I choose $D=\{ X_T \in A \}$ then clearly $D^c=\{ X_T \in A^c \} \in C$ but when D is of the form $D=\{ X_T \in A \} \cup \{T=\infty \}$ then $D^c=\{ X_T \in A^c\} \cap \{T< \infty \}$ which I cant show is in $\mathcal{C}$
Can you please help?
The statement of the problem is:
In the book, $X_T$ is defined only on $\{T<\infty\}$, so that $\{X_T\in A\}$ should be understood as $\{w\in\{T<\infty\}, X_T(w)\in A \}$.
$\Omega$ is in the collection: simply note that $\{X_T\in \mathbb R\}\cup \{T=\infty\} = \Omega$
The collection is closed under complementation:
$\{X_T\in A\}^c = \{T=\infty\}\cup \{X_T\in A^c\}$
$(\{X_T\in A\} \cup \{T=\infty\})^c = \{X_T\in A\}^c\cap \{T<\infty\} = \{X_T\in A^c\}$
The collection is closed under countable union: I leave it to you.