flangu schedules every review using a well-established body of research on how human memory works. This article explains the four ideas that underpin the system: the flashcard method, active recall, spaced repetition, and the FSRS-6 algorithm.
Flashcards
A flashcard is a question-and-answer pair. You see one side, try to produce the other, then check whether you were right.
That constraint matters. The card forces you to retrieve information, not just read it. Reading a word and its translation creates a weak memory trace. Recalling the translation from a prompt creates a much stronger one.
For vocabulary learning, flashcards are a natural fit. A word in the target language on the front; its translation, definition, and example sentences on the back. Each review is a small, focused retrieval attempt.
Active recall
Active recall is the practice of retrieving information from memory rather than reviewing it passively.
Passive study, such as re-reading notes or re-reading a list of vocabulary, feels productive but does little to strengthen memory. The act of retrieval itself is what consolidates a memory. This is called the testing effect, and it is one of the most replicated findings in cognitive psychology.
In flangu, every review is an active recall attempt. You see the front of a card and try to recall the back before tapping to reveal it. That moment of effortful retrieval is the learning event. Rating your recall honestly afterwards tells the system how well the retrieval succeeded, which affects when the card is shown again.
Tip
Rate cards accurately, not generously. The ratings are not a score. They are signals the algorithm uses to calibrate your schedule. Giving yourself Good when the answer took significant effort will cause the card to come back later than it should.
Spaced repetition
Spaced repetition is a study technique that schedules reviews at increasing intervals. The goal is to review a card just before you would forget it.
The underlying phenomenon is the forgetting curve: the probability that you can recall a piece of information drops rapidly after first learning it, then more slowly after each successful retrieval. Reviewing at the right moment flattens the curve. Each successful review extends the time before the next one is needed.
Practically, this means:
- A card you struggle with comes back the next day.
- A card you know well comes back after a week, then a month, then several months.
- A card you know very well may not need reviewing again for a year or more.
The interval between reviews grows only as fast as your recall justifies. That is what makes spaced repetition efficient: time is spent on cards you are at risk of forgetting, not on cards you already know.
Why this matters for vocabulary
Vocabulary retention is a pacing problem. Encountering a new word once or twice rarely leads to lasting recall. Seeing it dozens of times over a few days is also inefficient — the later reviews add little because you are still certain of the answer.
Spaced repetition solves this by distributing reviews across a timescale matched to the natural decay of memory. Words you find difficult are reinforced frequently. Words you know solidly are reviewed rarely, freeing time for new vocabulary.
The result is a gradually growing vocabulary where words are not just recognized briefly but retained over months and years.
The FSRS-6 algorithm
Older spaced repetition systems used fixed rules to grow intervals: multiplying the current interval by a constant after each successful review, for example. This works, but it applies the same schedule to every card and every learner.
FSRS (Free Spaced Repetition Scheduler) is a modern algorithm designed to be more precise. flangu uses FSRS-6, the current version.
How FSRS models memory
FSRS represents the state of each card in your memory using three values:
| Value | What it means |
|---|---|
| Retrievability (R) | The probability, right now, that you can recall this card. Decreases over time. |
| Stability (S) | How long it takes for R to drop from 100% to 90%, measured in days. A high S means a long interval before the next review. |
| Difficulty (D) | How hard it is to increase stability for this card. High-difficulty cards gain stability more slowly after each review. |
These three values are tracked independently for every card. A word you find easy and a word you find hard will have very different schedules, even if you first learned them on the same day.
What FSRS-6 does with your ratings
When you rate a card after a review, FSRS-6 updates that card's D, S, and R values using the outcome:
- Again: the card was not recalled. Stability is reduced significantly. The card is rescheduled for the near future.
- Hard: the card was recalled with effort. Stability grows slowly.
- Good: the card was recalled normally. Stability grows as expected.
- Easy: the card was recalled immediately. Stability grows more than normal.
FSRS-6 then calculates the next review date by finding the interval at which R is predicted to reach your desired retention level.
Desired retention
Desired retention is the probability of recall you want to maintain across your cards. The default is 90%, meaning flangu aims to schedule each card so that you have a 90% chance of recalling it at the next review.
A higher desired retention produces more frequent reviews. A lower value produces less frequent reviews but more forgetting. The 90% default is a reasonable balance for most learners.
Note
You do not need to adjust desired retention to use flangu effectively. The default value works well for the majority of learners.
Personalization
FSRS-6 learns from your review history. It uses machine learning to find the parameters that best describe the way you personally forget and retain information. Over time, as you complete more reviews, the schedule becomes better calibrated to your memory.
When you are new to flangu and have few reviews, FSRS-6 uses default parameters derived from the review history of tens of thousands of learners. Even with these defaults, the algorithm produces more efficient schedules than older fixed-interval methods.
What FSRS-6 does not do
FSRS-6 schedules reviews. It does not determine how long a session should last, how many new cards to introduce each day, or how to organize your decks. Those decisions remain yours. The algorithm works on whatever cards you put in front of it, as long as you review consistently and rate honestly.