Why this app?
This app's learning efficiency comes from combining three independently validated principles from learning science.

① FSRS (spaced repetition algorithm)
Each question carries a "stability" value, and the app resurfaces it right before predicted recall probability drops below a target retention threshold. Spaced repetition is repeatedly shown in cognitive psychology to produce significantly better long-term retention than massed practice (cramming) for the same number of reviews — the key point is that timing is decided mechanically from the theory, not by feel.

② Mastery-weighted question selection
A 0–100 mastery score updates based on correctness and answer speed, and lower-mastery questions get selected more often. This reflects the "desirable difficulty" principle: time is pulled away from already-mastered material and reallocated to weak points, raising net knowledge gained per unit of study time.

③ Focus-score-driven difficulty adjustment
Error rate and response-time variance are used to estimate real-time focus, which then adjusts difficulty and pacing. Under cognitive load theory, pushing hard questions while focus is low increases errors, and those error experiences can undermine retention or reinforce incorrect recall — so making difficulty adaptive to state is a rational way to prevent efficiency loss.

Role of gamification
RPG elements, ranks, and combo effects function as a separate motivational layer that keeps the person executing the plan above. However optimal a review schedule is in theory, it produces zero benefit if not carried out — so decoupling the retention algorithm from the engagement mechanic is logically sound design.

Conclusion
The app decides what, when, and at what difficulty to present material using evidence-based algorithms (FSRS, mastery weighting, focus adaptation) rather than intuition, and uses game mechanics to sustain execution of that plan. The exact efficiency gain can't be stated as a hard number, but each component implements independently-validated theory, making this design provably far more efficient than unstructured self-study.