Assume you have a system with unknown resistance $(R)$ at the design phase, therefore that could be modelled as a random variable, but time-invariant (i.e. the value of the random variable generated at time $t_0$ does not vary with time $t$, is constant over time, the stochastic process is stationary ergodic).
Now the load $(S)$ acting on the structure is also time-invariant. Load and resistance are independent variables.
In the design phase, the probability of failure at year 1 or at year 50 will always be the same, i.e. time-invariant. But what can be said about the probability of failure up to year 50? Is it the same as the probability of failure at year 1 or at year 50? Or it is given by:
$P_{f,50}=1-(1-P_{f,1})^{50}$