The learner desires less repetition and more memorization. However, the algorithm requires more repetition to adapt to the learner.
The learner wants their true retention to accurately match their desired retention. However, the algorithm requires data from a diverse range of retention levels to improve sampling efficiency and avoid regression.
The learner may choose their rating based on the interval length. For instance, if they find a review easy but see that the "easy" interval is too long, they might select "good" instead. They may only choose "easy" for very simple cards. This behavior leads to higher retention on those cards, causing the algorithm to adjust and lengthen intervals. It creates a vicious cycle.
