Most modern spaced repetition software is built on SuperMemo-2, the algorithm devised by Piotr Woźniak in the mid-1980s. At each review the user grades recall quality on a 0–5 scale: 0 for total blackout, 1 for incorrect but familiar, 2 for incorrect but easy once the answer is shown, 3 for correct with serious difficulty, 4 for correct after hesitation, and 5 for perfect recall. Grades below 3 count as failures: the repetition counter resets to zero, the next interval is set to one day, and the card's ease factor is reduced. Grades 3 through 5 are successes that move the card forward and may adjust its ease upward.
The interval schedule is straightforward. After the first successful review the card returns after one day; after the second success, six days; for the third and subsequent successes, the next interval equals the previous interval multiplied by the card's ease factor. The ease factor is a per-card multiplier with a default of 2.5 and a floor of 1.3; easier cards have their ease nudged up, harder cards down. Over time each card settles into its own rhythm - memory strength rather than a fixed timetable.
The Leitner system predates SM-2 and uses five (or so) discrete boxes: a correct answer moves a card up a box for a longer interval; an incorrect answer moves it back to box one for daily review. Its simplicity is appealing, but SM-2's continuous ease factor and six-grade scale allow finer-grained scheduling. Anki, the most widely used spaced repetition application, uses a modified SM-2 with two additional knobs - an "easy bonus" that lengthens the next interval when you press Easy, and a "hard interval" that shortens it when you press Hard.
In Anki these four answer buttons have distinct, non-interchangeable effects. Again fails the card and starts relearning with a short "learning step" while reducing ease. Hard repeats the card sooner than the default and slightly reduces ease. Good applies the standard SM-2 multiplier. Easy awards a bonus that grows the next interval beyond what the multiplier alone would give. Cards with an interval under 21 days are considered "young"; cards at 21 days or more are "mature", meaning they have been reliably learned but still need periodic review to remain stable.
Several deck options customize scheduling. Learning steps are the short review intervals (e.g., one minute, ten minutes) used during a card's initial learning phase; once the card clears them, it enters the review queue with the graduating interval, typically one day, then four days for a second graduation. The minimum interval, usually one day, prevents the scheduler from re-showing a card many times within a single session. The maximum interval, often 365 days, prevents a single card from disappearing for years and being forgotten in the meantime. The starting ease seeds new cards' initial ease factor; lowering it makes new cards harder from the start, raising it makes them gentler at the cost of more overdue cards later. A fuzz factor adds a small random jitter, typically ±5 percent, so that cards scheduled together don't all become due on the same day; an expansion multiplier around ×1.1 to ×1.3 trims the review load for long-interval cards by stretching them slightly outward.
A few habits keep the system healthy. Most adult learners can sustain about ten to thirty new cards per day per deck before backlogs compound and reviews spiral. A card flagged as a leech after many lapses (Anki's default threshold is eight) almost always means the card was poorly designed, not that your memory is faulty: rewrite it so it asks a single clean question, split it into smaller cards, or delete it. Studying cards from many topics in a single session, rather than blocking one topic at a time, beats blocked spacing for long-term retention. Varying the context of study - different examples, locations, or framings - builds richer retrieval routes and supports transfer to new situations.