A discrete event model was used to examine the effect of machine downtime and operating policy on the long-run average cost of an automotive stamping line. Operating policy refers to the selection of a target batch size and the circumstances under which a line stoppage will lead to the current batch being abandoned. It is assumed that the abandon/resume decision is based solely on the severity of the problem (ie repair cost) and the fraction of the batch completed. A method of identifying low cost operating policies is presented using data obtained from a real stamping plant. It is found that, within a single part framework, this approach results in significantly lower average costs than are currently achieved. It is also demonstrated that by varying the model parameters it is possible to measure the potential benefits arising from process modifications (eg decreased die-set times). This can be used to identify the areas where improvements will have the greatest impact on cost and is particularly useful when assessing the expected return on a potential investment. A multiple-part extension to the model is suggested and the potential benefits discussed.