We describe the effect of the equivalent pre-crack size (EPS) distribution errors on probability of failure (PF) of an FA/18 bulkhead. We distinguish two types of errors: (1) measurement errors of the EPS distribution parameters and (2) errors that arise because PF is typically calculated based on an EPS distribution with an infinite tail, while in a flying aircraft the tail cannot be unbounded. Our results indicate that both errors have a pronounced effect on accuracy of predicted PF. We also show that because predictions of the evolving crack sizes are less sensitive to the EPS distribution errors, they could be used as a complementary measure for estimating accumulated aircraft fatigue damage.