Advances in instructional technology have significantly increased the range of learning environments available to geographically distributed learners. Distributed combat readiness training need not be limited to individual skills now that students can gather in virtual spaces to train on collective skills critical to mission success. However, technological advancement has outpaced the capability to formally assess the quality of collective skills and to diagnose team performance deficits--a capability that is key to transforming learning environments into true instructional systems. In our research, we explored a range of methods for assessing the collective skills of U.S. Army National Guard officers in training as they war-gamed in a virtual tactical operations center. Our intent was to capture aspects of war-gaming performance that would (a) indicate the level of collective skill development; and (b) reveal the causes for team performance shortfalls. We based the design of our assessments on a psychological model of war-gaming that we developed using the results of a comprehensive cognitive task analysis. Our model comprises the psychological constructs associated with effective war-gaming processes and outcomes as well as with key determinants of war-gaming effectiveness. In collaboration with instructors, we administered our assessments to National Guard officers enrolled in the distance-learning version of the Armor Captains' Career Course, taught through the U.S. Army Armor School at Fort Knox, KY. Initial qualitative analysis suggests our war-gaming model and assessments are an appropriate and effective springboard for follow-on research. The primary lesson learned is that training-assessment validation efforts must involve an integrated approach in which assessment design is based on a thorough understanding of the training objectives, assessment implementation is based on a thorough understanding of the learning environment, and assessment validation is accomplished in a controlled setting in which specific hypotheses about the assessments can be tested.