FSRS vs SM-2: Why MindFlash Uses FSRS
MindFlash is an offline flashcard app that uses the FSRS spaced-repetition algorithm, not SM-2. This page describes the difference and why FSRS is the better choice for modern study.
The two algorithms
• SM-2 is the algorithm Anki has used since its launch. It was described by Piotr Wozniak in the late 1980s and is the de-facto standard in flashcard apps.
• FSRS (Free Spaced Repetition Scheduler) is a more modern algorithm developed by Jarrett Ye, based on the same underlying memory model as SM-2 but with a learning-to-rank approach that adapts to the user's recall history per-card.
Both algorithms are open-source. Both are based on the same "two components of memory" model: stability (how long a memory lasts) and retrievability (the probability of recall at a given moment).
What FSRS does better
FSRS's main advantage over SM-2 is that it learns from your review history. SM-2 uses a fixed difficulty parameter per card; FSRS adjusts the difficulty and stability based on how you actually recall each card.
Practically, this means:
• Cards you find easy are scheduled less often (less wasted review).
• Cards you find hard are scheduled more often (less forgetting).
• The model adapts to your study pattern, not the average study pattern.
The FSRS project is open-source at github.com/open-spaced-repetition. The algorithm has been validated in public benchmarks against SM-2 and shows fewer lapses at the same review count.
What FSRS does not do
FSRS is not magic. It does not make you remember cards you never study. It does not replace the need for daily review. It is a more modern algorithm that, for the same study time, produces better retention than SM-2.
Why MindFlash uses FSRS
SM-2 is the safer default. Anki, Quizlet, and most other flashcard apps use SM-2. FSRS is a deliberate upgrade — same time investment, better retention, no extra effort from the user.
If you have an Anki deck with a long review history, you can export the deck's review log and feed it to FSRS to retrain the model. MindFlash trains FSRS per-deck on your local review history automatically.
How MindFlash uses FSRS
Each deck in MindFlash has its own FSRS instance. The model trains on your review history for that deck and produces a per-card difficulty and stability. The daily review queue shows you the cards whose predicted retrievability is below your target threshold.
The threshold is configurable. The default targets high recall — the standard for "you actually remember this at exam time."
Get MindFlash
Free on Google Play. No account. No subscription. No in-app purchases.