Spaced repetition is underused for algorithm prep

The problem

You solve a medium-grade algorithm problem today. In one sitting, you internalize the pattern. Two weeks later, you attempt a near-identical problem and fail. The pattern was never consolidated into long-term memory.

This is the forgetting curve, first described by Ebbinghaus in 1885, validated across a century of cognition research. Initial encoding fades fast. Deliberate retrieval at increasing intervals is what pins it.

The standard spaced repetition schedule

Review the item at 1 day, 3 days, 1 week, 2 weeks, 1 month, 3 months. If you fail a review, reset to the start. This is the Anki algorithm, battle-tested across medical school, language learning, and professional certifications.

Why it maps to algorithm practice

An algorithm problem is an item. Solving it is initial encoding. Re-attempting from scratch after a gap is retrieval. If you succeed at the review, push the next review further. If you fail, reset to day one.

The retrieval must be cold. You cannot glance at the solution first. You attempt to re-derive it. This is the work that consolidates the pattern.

The schedule we generate internally

Day of first solve: mark the problem, record the pattern name, start the clock. Next day: cold re-attempt. Three days later: cold re-attempt. One week later: cold re-attempt. Two weeks, one month, three months.

If you fail at any review, reset the schedule but leave the problem in the pool. You will see it again at day 1 from that point. The next pass is easier because failure is itself a learning signal, often more lasting than success.

Why candidates skip this

Re-solving a problem you already know feels unproductive. The ego wants new problems to add to the tally. Spaced repetition requires the discipline to accept that consolidation beats accumulation.

Most prep platforms encourage accumulation because it is easier to gamify. New problems unlocked. Progress bar. Daily streak.

We instead weight retention. The algorithm is: show the user the review problems first, and only then unlock new ones. Two weeks in, the retention queue feels larger than the new queue. This is how it is supposed to feel.

Implementation today

Keep a plain text file. Three columns: slug, first solve date, next review date. Sort by next review date. Each morning, attempt the problems due today. Cold. Without looking.

If you succeed, multiply the gap by two. Standard multiplier is x2, but x2.5 gives slightly better retention in practice. If you fail, reset to day one.

That is a working spaced repetition system in about thirty lines of discipline. You do not need Anki, you do not need a product. You need to do the reps.

What Pro and Elite add

Automation of the schedule. Pattern-recognition so that failing a sliding-window problem also queues up related sliding-window problems for review. Leaderboard optional (we default to off because retention is between you and the work, not a social signal).