Flow is an intrinsically motivating (i.e. ‘autotelic’) psychological state of complete absorption in moment-to-moment activity that can occur when one performs a task whose demands match one's skill-level. Flow theory proposes that Flow causally leads to better performance, but empirical evidence for this assumption is mixed. Recent evidence suggests that self-reported Flow may not be linked to performance-levels per se, but instead to deviations from anticipated performance (the so-called flow deviation, or F~d effect). We aimed to replicate and extend these results by employing a high-speed steering game (CogCarSim) to elicit Flow, and specifically focused on the moderating effects of learning and task experience on the F~d effect. In a longitudinal design, 18 participants each played CogCarSim for 40 trials across eight sessions, totaling 720 measurements across participants. CogCarSim reliably elicited Flow, and learning to play the game fit well to a power-law model. We successfully replicated the F~d effect: self-reported Flow was much more strongly associated with deviation-from-expected performance than with objective performance levels. We also found that the F~d effect grew stronger with increasing task experience, thus demonstrating an effect of learning on Flow. We discuss the implications of our findings for contemporary theories of Flow.

心流是一种内在的激励（即“自发性”）心理状态，它完全沉浸在当下的活动中，当一个人执行一项其需求与自己的技能水平相匹配的任务时，就会发生这种情况。流理论提出，流会导致更好的性能，但这种假设的经验证据是混合的。最近的证据表明，自我报告的流量本身可能与绩效水平无关，而是与预期绩效的偏差（所谓的流量偏差，或F~d效应）。我们旨在通过采用高速转向游戏 (CogCarSim) 来引发 Flow 来复制和扩展这些结果，并特别关注学习和任务体验对F~d的调节作用影响。在纵向设计中，18 名参与者每人在 8 个会话中进行了 40 次试验，总共对参与者进行了 720 次测量。CogCarSim 可靠地引发了 Flow，并且学习玩游戏非常适合幂律模型。我们成功地复制了F~d效应：与客观性能水平相比，自我报告的 Flow 与偏离预期的性能更密切相关。我们还发现F~d效应随着任务经验的增加而增强，从而证明了学习对 Flow 的影响。我们讨论了我们的发现对当代心流理论的影响。