Following the evidence of research in critical thinking and cognitive bias training, we developed training designed to demonstrate people's decisions and actions affected by cognitive bias or elicit a bias in the player. Ideas were collected from and refined by experts with diverse backgrounds distributed geographically using an innovative solution. We then developed a video game that used multiple scenarios to teach cognitive bias recognition and mitigation. A learning and training effectiveness study indicated that the game was effective for learning although results were inconsistent for the different cognitive biases addressed by the game. The game recorded detailed data about scenes, scenarios and decisions each player made in the game. Called “breadcrumbsâ€�, this data detailed the path every player took. Traditional statistical analytic techniques tend to be clumsy instruments for breadcrumb analysis. Additionally, early aggregation and dimension reduction make the data more tractable but less meaningful. This paper details specific examples of how the breadcrumbs – paired with the study data – provided valuable answers in pinpointing areas to improve the game's learning effectiveness. The paper provides enough background information on the subject to enable the audience to appreciate the difficulty in cognitive bias training effectiveness and understand the examples shown. The majority of the paper discusses the data, the analysis, and the innovative data visualization techniques used. We discuss approaches that may prove more appropriate to extracting useful information from breadcrumb trails than traditional statistical analytic techniques. The audience will gain an understanding of the value of testing, data collection, and data visualization in training, education, simulations, and serious games. The paper will conclude with a discussion on using the techniques to improve the small batch testing in serious game development.