Educational video games can engage students in authentic STEM practices, which often involve visual representations. Specifically, because most interactions within video games are mediated through visual representations, video games provide opportunities for students to experience disciplinary practices with visuals. However, prior research has not investigated how students learn from visuals within games. Prior research on learning with visuals in non-game contexts suggests that visuals may confuse students if they lack prerequisite representational competencies which include sense-making competencies and perceptual fluency. Sense-making competencies allow students to relate visual features of a representation to the discipline-specific concepts they show and to explain connections between multiple visuals based on conceptual mappings. Perceptual fluency allows students to quickly and effortlessly extract relevant information from visuals and to fluently translate among different representational systems. To address this gap, we investigated the role of representational competencies for students' learning from educational video games. We conducted a 2x2 factor experiment with 120 participants to investigate the effects of sense-making and perceptual-fluency supports within the context of an educational video game. Results showed that sense-making supports did not enhance players' content learning. Further, perceptual-fluency supports enhanced players' content learning outcomes but only when they had high prior astronomy knowledge. Hence, interventions that support representational competencies in non-game environments may work differently in the context of educational video games. This suggests that designers of educational games may need to develop new strategies to support students' learning with disciplinary visual representations. [This paper will be published in "Computers & Education."]