spaced-repetition

Purpose

Schedule review of previously-learned concepts at expanding intervals to exploit the spacing effect. Manages due dates, adjusts intervals based on recall quality, and surfaces overdue items before they decay.

Activation

• Review session is due based on schedule. Learner initiates review. flashcards deck has due items. Sustained learning track (3+ sessions) needing interval maintenance.
• Skip if: one-shot session with no continuity. Concept not yet learned. Learner declines scheduling.
• Routing: overdue items take priority at session start. Coordinate with flashcards for card-level scheduling. Feed interval data to learning-memory.

Inputs

• Flashcard/item schedule with due dates, learner's current session, items from learning-memory.

Interval Algorithm (Simplified SM2)

Each item has an interval (days) and ease factor (EF, 1.3–2.5). Initial: 1 day → 3 days → then formula.

Score	Label	Interval Rule	EF Change
0	Failed	Reset to 1 day	EF -= 0.2 (min 1.3)
1	Hard	Stay at current	EF -= 0.1
2	Good	Interval × EF	No change
3	Easy	Interval × EF × 1.3	EF += 0.1 (max 2.5)

Mastery threshold: EF > 2.4, interval > 60 days, 5 consecutive successes → archive.

Workflow

1. Detect Due Items — Check next_review ≤ today. Sort by overdue duration (most overdue first). Report count.
2. Scope Session — ≤10 due: review all. >10: prioritize by overdue + weak-area overlap, cap at 15. Report deferrals.
3. Execute Review — Present front, wait for learner response, reveal back. Self-score: Failed/Hard/Good/Easy. Compute new interval immediately. Never reveal answer before attempt.
4. Handle Failures — Score 0: reset to 1 day, re-test at end of current session. Failed 3 sessions in a row: flag for misconception-detector.
5. Update Schedule — Output updated schedule. Show items due in next 7 days. Warn about upcoming review spikes.
6. Onboard New Items — Fresh concept → add at interval=1 day. Confirm concept is understood first (not still unclear).