In analyzing reliability data from repairable systems, engineers often will use the JMP Life Distribution platform to model the times between failures. However, such an analysis, which neglects the time order of the observations, can lead to incorrect conclusions. The Recurrence Analysis platform in JMP provides techniques for correctly modeling reliability data from repairable systems. This talk will include examples of improper application of the life distribution approach to reliability data and the discovery analysis resulting when the appropriate recurrence analysis methodology is applied. The talk will also highlight the limitations of mean time between failure (MTBF) as a reliability summary statistic and illustrate how MTBF values reported in the literature can be artificially inflated.