FSRS vs SM-2: Which Spaced Repetition Algorithm Wins in 2026?

If you are comparing FSRS vs SM-2, the short answer is: FSRS is the more accurate, adaptive algorithm for most long-term learners, while SM-2 remains a simple, battle-tested option for smaller decks or users who prefer predictable heuristics.

Both algorithms schedule flashcard reviews, but they make predictions in fundamentally different ways. FSRS uses a machine-learning memory model trained on your personal review history. SM-2 relies on fixed multiplication rules that have barely changed since 1987. That difference becomes meaningful when your deck grows past a few hundred cards.

This guide breaks down how each algorithm works, where FSRS wins, when SM-2 still makes sense, and how to choose the right scheduler for your workflow.

What is FSRS?

FSRS stands for Free Spaced Repetition Scheduler. It is an open-source spaced repetition algorithm created by Jarrett Ye (L-M-Sherlock) and released in 2022. FSRS is built on the DSR memory model: Difficulty, Stability, and Retrievability.

After every review, FSRS updates the Difficulty and Stability values for that card. It then schedules the next review so that Retrievability hits your chosen desired retention — usually 90–92% — at the exact moment you see the card again.

FSRS has been integrated into Anki since version 23.10 and is the default scheduler for new decks in recent releases. It is also used natively by apps like Deckbase, Mochi, and RemNote.

What is SM-2?

SM-2 is the SuperMemo-2 algorithm developed by Piotr Wozniak and published in 1987. It became famous because Anki adapted a modified version of it as its default scheduler for many years.

SM-2 works with a single value called the ease factor. Each time you rate a card, the algorithm multiplies the current interval by the ease factor to produce the next interval. Hard ratings lower the ease factor; easy ratings raise it.

The algorithm is simple and robust. It has helped millions of learners memorize vocabulary, medical facts, legal codes, and technical concepts. However, it does not model memory directly. It uses fixed rules that do not adapt to individual forgetting curves beyond the ease factor.

FSRS vs SM-2: side-by-side comparison

The biggest practical difference is personalization. SM-2 applies the same formula to every user. FSRS learns from your actual ratings and optimizes intervals for your memory.

Prediction accuracy: what the benchmarks say

The most widely cited comparison comes from the Expertium benchmark, which evaluated schedulers against more than 700 million Anki reviews.

Lower numbers mean the algorithm is better at predicting whether you will remember or forget a card. FSRS-5 cuts prediction error by roughly two-thirds compared with SM-2. In real study sessions, that translates to fewer surprise lapses and fewer unnecessary reviews of material you already know.

When FSRS is the better choice

FSRS tends to win for learners who:

• Have decks with 500+ cards
• Study for months or years, not days
• Want to set an explicit retention target
• Mix easy and hard material in the same deck
• Already have 1,000+ reviews of history for parameter optimization
• Want to minimize daily review time without sacrificing recall

Medical students, language learners, and certification candidates often see the biggest gains because their study timelines are long and their card libraries are large.

When SM-2 still makes sense

SM-2 is still a reasonable choice if you:

• Run a small or short-lived deck (under 500 cards)
• Are new to spaced repetition and want a simple scheduler
• Prefer predictable, manual control over algorithmic optimization
• Have a legacy Anki setup that already works
• Do not yet have enough review history for FSRS to train on

SM-2 is not broken. For many users, the difference between SM-2 and FSRS is smaller than the difference made by card quality, consistency, and honest self-rating.

How to switch from SM-2 to FSRS

In Anki 23.10 or later:

1. Open a deck and click Options.
2. Scroll to the Advanced section.
3. Enable FSRS.
4. Set your desired retention (start at 90–92%).
5. Click Optimize to fit parameters to your review history.

The switch is non-destructive. Your cards and review history stay in place; only future intervals are scheduled differently.

In Deckbase, FSRS is the built-in scheduler. There is no toggle — every review uses FSRS automatically. If you import an Anki deck via APKG, your cards transfer, but scheduling history does not. You start fresh with FSRS intervals from day one.

A practical 14-day test

If you are unsure which algorithm to use, run this two-week experiment:

1. Pick one active deck with at least 200 cards.
2. Enable FSRS and set retention to 90%.
3. Track daily review time, completion rate, and lapse count for 14 days.
4. Compare with your last two weeks on SM-2.
5. Decide based on outcomes, not theory.

Most users who complete this test stick with FSRS because the daily workload drops while retention stays stable.

FAQ

Is FSRS better than SM-2?

For most long-term learners with large decks, yes. FSRS predicts forgetting more accurately and can reduce review volume by 20–30% at the same retention level. SM-2 is still fine for small decks or users who prefer simplicity.

Does FSRS work without much review history?

FSRS uses default parameters until it has enough data to optimize. You generally need around 1,000 reviews before personalized parameters become reliable. Until then, FSRS still performs at least as well as SM-2 for most users.

Can I switch back from FSRS to SM-2?

Yes. In Anki, you can disable FSRS in deck options and return to SM-2. Your future intervals will revert to SM-2 scheduling, but past review history is preserved.

Does Deckbase use FSRS or SM-2?

Deckbase uses FSRS natively as its core scheduler. There is no SM-2 option. This keeps the app simple and ensures every user benefits from modern retention prediction without manual configuration.

What retention target should I set?

Start at 90–92%. Lower targets reduce workload but increase lapses. Higher targets improve recall but can make daily reviews unsustainable. Adjust after 2–3 weeks based on your actual completion rate and study time.

Final verdict

Choose FSRS if you want a modern, adaptive scheduler that reduces review load and improves prediction accuracy for long-term learning.

Choose SM-2 if you have a small deck, prefer simplicity, or already have a workflow that works and do not want to change it.

For most serious learners in 2026, FSRS is the clear default. The gap is not just theoretical — it shows up as shorter daily sessions, fewer surprise lapses, and more stable long-term retention.

Ready to study with FSRS? Try Deckbase(opens in a new tab) and review with the same algorithm Anki uses — without the setup complexity.