The U.S. Army continues to employ constructive and game-based simulation for training. While both classes of simulation have been found to lower the cost of training, it is still unknown whether or not these classes are actually effective training mechanisms. The Linguistic Geometry Real-time Adversarial Intelligence and Decision-making (LG-RAID) simulation is a lightweight, game-based, constructive simulation that exploits novel game theory to create intelligent, predictive and tactically-correct Courses of Action (COAs) for exercise participants at the company echelon and below. The primary goal of this study was to examine the training effectiveness and usability of the U.S. Army's LG-RAID simulation in an operationally relevant environment. A secondary objective of this study was to assess both the usability and functionality of the simulation in order to improve the technology through future design recommendations. Qualified Soldiers were randomly assigned to one of two training treatments (LG-RAID or a traditional planning method) and tasked to develop, plan and brief a tactically sound operational mission in order to empirically assess the training effectiveness of LG-RAID. The independent variable was training treatment. Dependent variables included performance and individual survey responses. Experimentation was conducted at Fort Benning, GA and performance was evaluated by accredited Army instructors. Results of this study indicate that LG-RAID shows promise as an effective training simulation tool when compared to the baseline condition.