Analysis of repair data

A simple, informative plot of data on numbers and costs of repairs on a sample of systems is presented. The plot provides a nonparametric estimate of the mean cumulative (1) cost or (2) number of repairs per system versus system age. The plot is used to: (1) evaluate whether a cost (or repair) rate increases or decreases with age, which helps one decide on system retirement and burn-in; (2) compare different samples (from different designs, production periods, operating environments, maintenance policies, etc.); (3) predict future numbers and costs of repairs; and (4) reveal unexpected information and insight an advantage of plots. The plot also applies to data on recurrent disease episodes, factory simulation data, borrower credit, and other histories