Spaced repetition¶

Spaced repetition is a study technique that schedules reviews at increasing intervals. The core insight: reviewing something just before you forget it is more efficient than reviewing it on a fixed schedule. Problems you know well are shown infrequently; problems you struggle with appear often until they stick.

hone uses a modified SM-2 algorithm, the same algorithm behind Anki and many other SRS tools.

The SM-2 algorithm¶

Each problem has an easiness factor (EF), an interval (days until next review), and a repetition count (consecutive successes).

After every attempt, hone:

1. Derives a quality score (1–5) from the result and solve time
2. Updates the EF using the SM-2 formula
3. Computes a new interval based on the repetition count and EF
4. Schedules next_review_date = today + interval

Quality scoring¶

Quality is never self-reported. hone derives it automatically:

Thresholds are per-difficulty and configurable (see Configuration).

Interval progression¶

On each consecutive success:

On failure, the interval resets to 1 day and the repetition count resets to 0.

Easiness factor¶

The EF starts at 2.5 and is updated after every attempt:

EF' = EF + (0.1 - (5 - q) × (0.08 + (5 - q) × 0.02))

Where q is the quality score (1–5). The EF is clamped to a minimum of 1.3 — problems never schedule faster than approximately weekly at steady state.

A higher EF means longer intervals (the problem is "easy" for you). A lower EF means shorter intervals (needs more practice).

Mastered problems¶

When a problem's interval exceeds 60 days, it's marked mastered_before = true. This is a one-way flag — once mastered, always considered mastered for stats purposes.

If a mastered problem is failed (EF decays, interval resets), it resurfaces for re-practice. Once you succeed on it twice in a row again, hone recognizes you still know it and applies an interval boost (doubling the computed interval), so it doesn't clog your queue for months.

The picker¶

hone practice selects the next problem using:

1. Filter — if a playlist or topic is active, only consider those problems
2. Due problems — all problems where next_review_date ≤ today
3. Pick rule — if due problems exist, pick the most overdue (oldest date first); ties break by difficulty (easy first), then playlist position or random. Otherwise pick the one with the nearest upcoming date

The TUI makes the distinction clear: "you have 5 problems due" vs "nothing due today — here's one in 3 days."

New problems¶

When a problem is first added, it gets:

• easiness_factor = 2.5
• interval_days = 1
• repetition_count = 0
• next_review_date = today

It's immediately due for review.