I have out of time validation sample data but, how do I measure the performance of the model?

Advantages, if any, of deadly military training? Other fitters (e.g. Actually, it looks like having a higher monthly charge means you’re more likely to stick around. Who is the "young student" André Weil is referring to in his letter from the prison? Do I still need a resistor in this LED series design? But, if you have $X(t)$, then you must be making measurements on a customer, so they are not dead! Why didn't the Imperial fleet detect the Millennium Falcon on the back of the star destroyer? Sorry to bother you again. The model object CoxTimeVaryingFitter() currently does not support or include functions to predict survival probability directly. MathJax reference. Simply put: you can't predict for epistemological reasons. How do you win a simulated dogfight/Air-to-Air engagement?
Therefore, the fact that you have or don't have $X(t)$ tells you if the customer is alive or not, making any "survival" prediction moot. Does this the below approach makes sense?