This study investigated pedagogical agent (with/without) and feedback (explanatory/confirmatory) types on achievement, flow experience, and cognitive load in a digital game-based learning environment. We developed a game about spreadsheets and prepared four versions of the game in the context of pedagogical agent and feedback types. Through the mixed-methods sequential exploratory design, we applied game versions in four experimental groups (154 undergraduate students) within a two-week spreadsheet subject. We collected quantitative (achievement test, flow experience scale, cognitive load scale, elapsed time, skill, and challenge-level questions) and qualitative (semi-structured interview technique and open-ended interview form) data. According to the results, the achievement scores of all participants increased regardless of the pedagogical agent and feedback type. There was no significant difference in cognitive load, flow experience, skill level, or challenge level according to the pedagogical agent and feedback types. Confirmatory feedback groups spent more time than explanatory feedback groups. The dependent variables had significant relationships. Students expressed that game-based learning promoted improvements in the learning process.