FSRS Algorithm
How Neurako's FSRS scheduler models your memory using stability, difficulty, and retrievability to schedule reviews at the right time.
Last updated 2026-05-23
FSRS (Free Spaced Repetition Scheduler) is the algorithm Neurako uses to schedule every review. It tracks three memory properties per card — stability, difficulty, and retrievability — and uses them to find the interval that keeps recall just at your target probability. The result is a schedule that adapts to your actual memory, not a generic average.
Key Takeaways
- Stability determines how long until a card needs review; difficulty controls how fast stability grows - The default target recall probability is 90% — FSRS schedules each card's review for when retrievability is about to drop below that - Rating a card Again resets stability significantly; rating Easy grows stability faster and slightly reduces difficulty - Pro users can adjust desired retention, maximum interval, and enable FSRS weight personalization based on their review history
The three FSRS memory variables
FSRS models memory with three values that are tracked independently for every card in your collection.
Stability (S)
Stability is the number of days until the card's recall probability naturally drops to 90%. A card with a stability of 7 will be due roughly 7 days after a successful review. A card with a stability of 180 won't need review for about six months.
After a Good or Easy rating, stability increases — sometimes substantially. After an Again rating, stability resets to a lower post-lapse value (called the "lapse stability"), which is still higher than a brand-new card but much lower than the pre-lapse value. This captures the real memory phenomenon that a forgotten card is easier to relearn than to learn for the first time.
Difficulty (D)
Difficulty sits on a 1–10 scale and represents how resistant a card is to improvement. High-difficulty cards (closer to 10) grow stability more slowly after successful reviews. Low-difficulty cards (closer to 1) grow stability quickly.
Difficulty is initialized based on your first rating and updated after each review. Rating Hard increases difficulty; rating Easy decreases it. This means a card you consistently struggle with will gradually become harder to push out to long intervals — accurately modeling the reality that some material is inherently harder to retain.
Retrievability (R)
Retrievability is your estimated probability of recalling the card right now, at this moment. Immediately after a review, R is 1.0 (100%). Over time, as memory naturally fades, R decreases according to the FSRS forgetting curve formula.
The scheduler uses R continuously: a card is due when R falls to the target retention level (90% by default). This is why interval lengths vary — a card with high stability might take weeks to drop to 90% retrievability, while a high-difficulty card might drop there in a day or two.
How scheduling works
When you rate a card, FSRS:
- Updates the card's stability based on the rating, the current difficulty, and the elapsed time since the last review.
- Recalculates difficulty based on the rating.
- Finds the interval
twhere the forgetting curve predicts R will equal the desired retention target. - Schedules the next review for today +
tdays.
The default desired retention is 0.9 (90%). This means Neurako aims to present each card for review just before you have a 1-in-10 chance of having forgotten it. Most of your reviews will be successful at this target — the Again rate for healthy review queues tends to be 10–15%.
FSRS Visualizer
Adjust a card's state and see the next interval move
Rating
Projected next review
21d
This is a simplified illustration of the same intuition FSRS uses: stable cards can be pushed out, difficult cards need shorter loops, and honest ratings matter more than aggressive intervals.
Fuzz
FSRS applies a small random jitter (called "fuzz") to calculated intervals. A card that would naturally be due in exactly 14 days might be scheduled for 13 or 15 instead. This prevents large batches of cards from clustering on the same review day — a common problem with systems that use fixed intervals for many cards created at the same time.
Fuzz is enabled by default and does not meaningfully affect long-term learning outcomes.
What each rating does
| Rating | Meaning | Effect on stability | Effect on difficulty |
|---|---|---|---|
| Again (1) | Forgot the card | Significant drop to lapse stability | Increases |
| Hard (2) | Correct but struggled | Smaller increase than Good | Increases slightly |
| Good (3) | Correct with normal effort | Normal increase per the model | Minor change |
| Easy (4) | Instant recall | Larger increase than Good | Decreases |
Rating honestly is the most important thing you can do to keep FSRS accurate. If you rate Easy on cards you aren't fully confident on, stability grows faster than your memory supports, and you'll see more lapses in the future.
Why FSRS beats fixed-interval systems
Older spaced repetition systems (including SM-2, which Anki historically used) scheduled reviews based on fixed multipliers. Every card used the same formula regardless of how that specific card behaved in your memory. FSRS instead fits a model to each card's actual review history.
The practical result: cards you know well move to very long intervals quickly. Cards you struggle with stay in shorter cycles until they genuinely stabilize. The schedule reflects your real memory, not a statistical average across all users.
For a detailed comparison, see FSRS vs SM-2.
FSRS personalization
FSRS personalization is available on the Pro plan. Go to Settings → Study to access the following options.
Desired retention — Adjustable between roughly 70% and 97%. A higher target (e.g., 95%) means cards come back more frequently; a lower target (e.g., 80%) allows longer intervals at the cost of a somewhat higher forgotten-card rate. The default 90% is a well-calibrated balance for most learners.
Maximum interval — Caps the longest interval FSRS can schedule. If you're studying for an exam in 3 months, you might cap intervals at 90 days to ensure everything is reviewed at least once before the exam.
FSRS weight optimization — Neurako can analyze your review history and tune the FSRS model parameters to better match your personal memory characteristics. This is an advanced option that improves scheduling accuracy for users with a substantial review history.
Spaced Repetition in Neurako
How the day-to-day review experience works — card states, rating buttons, and cram mode.
FSRS vs SM-2
A technical comparison of FSRS and the older SM-2 algorithm.
Spaced Repetition Explained
The memory science behind why spaced repetition works.
Related journal guide
For a more opinionated product comparison or narrative walkthrough, read FSRS explained for learners on the Neurako Journal.
Ready to turn this into a study ritual?
Start studying with NeurakoRelated reading
FSRS vs SM-2
A technical and practical comparison of the SM-2 and FSRS spaced repetition algorithms — what changed, what it means for learners, and why Neurako uses FSRS.
Spaced Repetition Explained
How spaced repetition works, why timing matters, and how adaptive schedulers make reviews more efficient.
Spaced Repetition in Neurako
How the review queue works in Neurako — card states, rating buttons, interval previews, cram mode, and daily study habits.