Spaced Repetition for LeetCode

You can solve 300 LeetCode problems and still blank in the interview — because solving a problem once does not mean you will remember it six weeks later. Spaced repetition fixes that. Here is the method, the science, and how to run it without hand-building flashcards.

What spaced repetition is

Spaced repetition schedules each review for the moment you are about to forget. Get a problem right and the next review pushes further out; get it wrong and it snaps back to soon. Over a few cycles the interval stretches from a day to weeks to months, and the pattern lands in long-term memory for a fraction of the total study time.

Why it beats grinding new problems

Grinding rewards volume, not retention. The 250th new problem you have never reviewed is worth less than the 50 you can re-derive cold. Interviews test whether you can produce a pattern under pressure, not whether you once saw it. Spaced repetition optimizes for exactly that.

The two principles that matter

Two ideas do the heavy lifting. Active recall: do not re-read your old solution — try to reproduce it from a blank editor first. The struggle is the learning. Spacing: revisit on an expanding schedule rather than cramming. Cramming gets you through tomorrow; spacing gets you through the interview.

How to run it for interviews

• Solve a problem on LeetCode.
• Log it with the key idea — the pattern, not the full code.
• When it comes due, re-derive it from scratch before looking.
• Rate how it went: a clean solve stretches the interval, a fail resets it.
• Watch your clean-rate per pattern climb over the weeks.

Doing it without flashcard busywork

The friction in spaced repetition is building and maintaining the cards. AlgoLoop removes it: a browser extension captures each LeetCode solve automatically, categorizes the pattern, hides your old solution until you re-attempt, and schedules every review for you. You solve; it remembers what you need to revisit.

Common questions

How many problems should I review a day?

Set a daily cap — 15 to 25 reviews is sustainable for most people. AlgoLoop spreads a backlog under your cap so you never get a 60-problem wall on one day.

Is spaced repetition better than just doing more problems?

For retention, yes. New problems build coverage; spaced review builds recall. You need both, but most people over-grind new problems and never revisit, then forget under interview pressure.

Do I have to build flashcards?

No. AlgoLoop captures each solve from the browser automatically and schedules the review, so there is no manual card-building — the usual reason people quit spaced repetition for LeetCode.