Spaced repetition shows each flashcard at a growing interval timed to just before you would forget it, which is the most efficient way to commit facts to long-term memory. This calculator runs the exact SM-2 algorithm that powers Anki so you can see how a single review reschedules a card.
How it works
SM-2 tracks three values per card: the repetition count n, the ease factor EF,
and the interval I in days. After grading your recall quality q from 0 to 5,
the new ease factor is:
EF' = EF + (0.1 - (5 - q) * (0.08 + (5 - q) * 0.02))
EF' = max(EF', 1.3)The next interval is then chosen by repetition number:
if q < 3: n = 0, I = 1 (failed — restart)
if n == 1: I = 1
if n == 2: I = 6
if n >= 3: I = round(previous I * EF')The quality scale — what each grade means
The q value (0–5) is the key input that makes SM-2 adaptive. Many people grade too generously (“I sort of knew it — that’s a 4”), which gradually inflates intervals beyond the forgetting curve:
| Grade | Label | Meaning |
|---|---|---|
| 0 | Blackout | No memory of the card whatsoever |
| 1 | Incorrect, serious difficulty | Wrong answer, hard even seeing the correct one |
| 2 | Incorrect, but close | Wrong, but the correct answer felt familiar |
| 3 | Correct, significant effort | Right answer, but only after a real struggle |
| 4 | Correct, minor hesitation | Right answer with a brief pause or slight uncertainty |
| 5 | Perfect recall | Immediate, confident, effortless recall |
Any grade below 3 resets the repetition counter and sends the card back to day 1. Any grade of 3 or above counts as a pass and advances the interval, but grades below 5 reduce the ease factor slightly.
How the ease factor evolves over time
The ease factor starts at 2.5 for a new card (the Anki default). Each review adjusts it by a small amount based on the quality grade. The update formula means:
- A grade of 5 (perfect) increases EF slightly:
EF' = EF + 0.1 - A grade of 4 leaves EF nearly unchanged:
EF' = EF - A grade of 3 reduces EF slightly:
EF' = EF - 0.14 - A grade of 2 reduces EF more:
EF' = EF - 0.32
The minimum ease factor is 1.3, which prevents a difficult card from being reviewed every single day indefinitely. A card that consistently receives grade 3 will have its ease factor gradually reduced, meaning it gets reviewed more often than an easy card with a higher EF.
Worked example: third review
A card on its third review with ease factor 2.5 and a previous interval of 6 days:
- Grade 5 (perfect): new EF = 2.6, next interval =
round(6 × 2.6) = 16days - Grade 4 (minor hesitation): new EF = 2.5, next interval =
round(6 × 2.5) = 15days - Grade 3 (struggled): new EF ≈ 2.36, next interval =
round(6 × 2.36) = 14days - Grade 2 (wrong): n resets to 0, new EF ≈ 2.18, next interval = 1 day
Practical advice for better results
Be honest with grades. Rating cards at 4 when it was actually a 3 feels kind, but it inflates intervals and means you will forget the card at the next review. The algorithm is only as good as your self-assessment.
Use it for pure memorisation tasks. SM-2 is well-suited for vocabulary, formulas, historical dates, medical terminology, and other items where recall is binary (you know it or you don’t). It is less suited for procedural skills or conceptual understanding that requires application rather than retrieval.
Consider suspend rather than grade 0. Cards you fail repeatedly because they are poorly phrased or too complex should be edited or split. Continuously failing a card just pushes it to appear every day without building retention — fix the card first.
The first two repetitions are fixed. The initial intervals (1 day, then 6 days) are hardcoded in SM-2 regardless of ease factor. Interval multiplication only begins from the third repetition onward.