This paper addresses the question of how free riding in large-value payment systems should be properly measured. Based on the valuable proposal by Denbee, Garratt and Zimmermann, various measures of free riding in large-value payment systems are investigated. Assessing the measures proposed in the literature against a list of rational postulates for measurement reveals their weaknesses. To overcome these weaknesses, we define three other measures, which are designed to be used for econometric studies and yield results independent of the size, composition and other special features of the payment system. Empirical results for nine important participants in TARGET2-BBK are displayed and compared. It turns out that a combination of at least two measures would be recommended for capturing the various aspects of free riding. The measures reveal some stable payment behavior for most banks over time, but also some remarkable regime shifts that yield interesting insights about single participants. The calculated levels of free riding can be judged to be rather low given the empirical results. In addition, the stable pattern over time raises the question of whether other unobserved features explain the different payment patterns such that labeling as a free rider based purely on measurement would be premature.