Smart CAT (Smart Configurable Assessment Tool) is a software for assessments in serious games, As such, it relies on the principles of an unobtrusive formative assessment methodology, namely stealth assessment (SA).

SA requires specifying competency models, which are tree-like structures describing competencies and their constituents. Given that certain statistical assumptions hold, it allows the coupling of these competency models with gameplay data to form statistical models. In turn, these statistical models can be represented in machine learning (ML) algorithms in order to generate inferences regarding learners’ mastery level.

Although SA is a proven methodology, its application is a complex, laborious, and time-consuming process that requires hardcoded solutions. To lower these barriers, Smart CAT comes as a stand-alone software solution which detaches SA from specific games’ source code implementations and allows its end-users to

• Easily configure statistical models,
• Optimize the ML processing of the learner data, and
• Examine the validity and reliability of their assessments.