Objective: The goal of this study was to evaluate two kinds of difficulty adaptation techniques in terms of enjoyment and performance in a simple memory training game: one based on difficulty-performance matching (“task-guided”) and the other based on providing a high degree of control/choice (“user-guided”). Methods: Performance and enjoyment are both critical in making serious games effective. Therefore the adaptations were based on two different approaches that are used to sustain performance and enjoyment in serious games: 1) adapting task difficulty to match user performance by leveraging the theories of zone of proximal development and flow, thus maximizing performance that can then lead to increased enjoyment and 2) providing a high degree of control and choice by using constructs from self-determination theory, which maximizes enjoyment, that can potentially increase performance. 24 participants played a simple memory training serious game in a fully randomized, repeated measures design. The primary outcome measures were enjoyment and performance. Results: Enjoyment was significantly greater in user-guided (p < 0.05), whereas performance was significantly greater in task-guided (p < 0.05). Conclusion: The results suggest that a trade-off between maximizing performance and maximizing enjoyment could be achieved by combining the two approaches into a “hybrid” adaptation mode that gives users a high degree of control in setting difficulty, but also advises them about optimizing performance.