As weapon systems have become more complex, the ability to discern performance by simply building systems correctly and placing them in operational use has diminished greatly. This complexity arises from the interaction of more components and subsystems. People are generally poor at predicting patterns formed from the interactions of many elements (e.g., rules and computing artifacts). The only way that we may fairly elicit patterns of performance over the full range in which we want these complex systems to perform without introducing bias is to develop useful metrics and apply the scientific method to understand operational, technical, and systemic interactions. This is the focus of this chapter.