Fibonacci intervals: 1 → 2 → 3 → 5 → 8 → 13 → 21 → 34 → 55 days. Built from data/progress.txt, the daily practice log.
Loading practice log…
A session, not a backlog. Every overdue problem is scored by how far past its interval it is × how much it is worth (MUST, Blind 75, NeetCode, Top 100 Liked, difficulty) × how often it has come back marked “again”, and the slots are then spent one topic at a time, starting with whichever topic is getting the least practice for its weight — so a good week on DP cannot swallow the whole list. Same scoring as suggest-review, the command-line planner.
Importance is each topic's share of the whole README — every problem it holds, weighted by MUST, the curated lists and difficulty. Attention is its share of the practice actually logged in the window below, weighted so that yesterday counts for more than three weeks ago, and counting sessions rather than distinct problems: grinding one problem five times is five units spent on that topic. A ratio near 1.0 means a topic is getting its fair share.
| Topic | Problems | Importance | Attention | Ratio | Last practised |
|---|
The command-line version reads git history as well, and can write the plan to markdown: suggest-review.
| Problem ↕ | Topic ↕ | First seen ↕ | Last seen ↕ | Times ↕ | Next review ↕ | Again ↕ | Status |
|---|