FSRS explained: the modern spaced repetition algorithm

What FSRS stands for

FSRS — the Free Spaced Repetition Scheduler — is an open-source algorithm that decides when you should next see each thing you’re learning. It combines a model of human memory with machine learning, fitting itself to your actual review history rather than applying one fixed formula to everyone.

FSRS vs SM-2 (the old default)

For decades, most flashcard apps used SM-2, the SuperMemo 2 algorithm. SM-2 is battle-tested but rigid: it gives every learner the same intervals and can drift into “ease hell,” scheduling more reviews than you actually need. FSRS instead predicts, for each individual card, the probability that you’ll still remember it — and schedules the review for the moment that probability drops to your target.

Benchmarks on hundreds of millions of reviews show FSRS needs roughly 20–30% fewer reviews than SM-2 for the same retention.

Fewer reviews for the same memory means either less time per day, or more words learned in the time you have. Missing a card also no longer resets you to day one, which takes a lot of the frustration out of daily review.

Why it’s trusted

FSRS is open-source, independently benchmarked, and now built into Anki, the most widely used flashcard tool in the world. It’s not a proprietary black box — it’s a community-vetted standard.

How 10 Words uses FSRS

In many apps you have to enable and tune FSRS yourself. In 10 Words it’s the default, working out of the box: every word you learn is scheduled by FSRS from the first day, so your ten-words-at-a-time habit is always reviewing exactly what’s about to slip. Pair that with learning in context and you get efficient and durable.