Spaced Repetition Algorithms: SuperMemo SM-2 vs. SM-17

Dr. Piotr Wozniak devoted decades to modeling memory decay mathematically. His second digital attempt, the **SM-2** algorithm, serves as the engine for platforms like Anki and Polyvo. However, years of learning data eventually birthed the highly sophisticated **SM-17** algorithm.

Limitations of the SM-2 Model

SM-2 schedules intervals using fixed algebraic steps based on instant user quality ratings. Its Easiness Factor (EF) lacks memory of historical patterns, occasionally leading to redundant card reviews or unexpected forgetting spikes.

The Modern SM-17 Paradigm

SM-17 represents memory status as a coordinate inside a three-dimensional space: Retrievability, Stability, and Difficulty. By fitting individual response histories to global datasets, it forecasts memory decay rates with far greater precision.