Methods · Interactive Tool
A scheduler that shows its working, not just an answer.
This tool runs Piotr Woźniak's SM-2 algorithm — published in full, exact detail in 1990 — entirely in your browser. Tell it what you're learning and how your last review went; it computes the next interval and ease factor using the same arithmetic Woźniak's original SuperMemo program used, and shows every number along the way. Nothing is sent anywhere.
SM-2 (Woźniak, 1990) — the version this tool reproduces exactly, formula for formula. See Sources.
01 — The Algorithm
Six numbers, one formula.
Every item you're tracking has three numbers behind it: n (how many times in a row you've recalled it correctly), EF (an "ease factor" starting at 2.50, tracking how easy this item is for you specifically), and I (the interval in days until the next review). After each review you grade your own recall on a 0–5 scale, and the algorithm updates all three:
- If you scored 3 or above (a real, if imperfect, recall): the first interval is 1 day, the second is 6 days, and every one after that is the previous interval multiplied by the ease factor, rounded up.
- If you scored below 3 (you didn't really get it): the count resets to zero and the item is treated as new again — back to a 1-day interval — though its ease factor is still updated by the same formula, not reset.
- The ease factor itself moves by
EF' = EF + (0.1 − (5−q)×(0.08+(5−q)×0.02)), floored at 1.3 so no item's interval can shrink below usefulness.
Woźniak tuned those specific constants (0.1, 0.08, 0.02, the 1.3 floor, the 6-day second step) by trial and error on his own vocabulary learning — not a universal law of memory, but a formula precise and published enough to implement exactly, which is why it's used here instead of a vaguer "review more often" rule. In his own account of the first year running it: 10,255 items memorised, 41 minutes a day, 89.3% overall retention (92% once very recently learned items are excluded).
| Grade | Meaning |
|---|---|
| 5 | Perfect response |
| 4 | Correct response after a hesitation |
| 3 | Correct response, recalled with serious difficulty |
| 2 | Incorrect response; the correct one seemed easy to recall once shown |
| 1 | Incorrect response; the correct one remembered once shown |
| 0 | Complete blackout, no recognition at all |
02 — The Tool
Build your schedule.
Add what you're learning, say how your last review went, and this computes the next one. Add several items to build a real schedule.
| Item | Reviews (n) | Ease (EF) | Interval | Next review | Grade it now |
|---|
Grading a row in the table simulates the next review landing exactly on its scheduled date, so you can click forward through several repetitions at once. Nothing here is saved — reload the page and the schedule clears.
03 — Worked Example
If every review were graded a perfect 5
Computed live by the same function as the tool above, starting from a brand-new item (n=0, EF=2.50) reviewed on 26 Jul 2026 — not typed in by hand, so there's no arithmetic slip between the demo and the real tool.
| Repetition | Grade | Ease before → after | Interval | Review date |
|---|---|---|---|---|
| Calculating… | ||||
Notice the interval grows faster than the ease factor alone would suggest, because each interval multiplies the previous interval, not the original one — this compounding is the entire mechanism behind spacing that stretches from days to months.
04 — Common Questions
Before you trust the dates.
Is SM-2 the "best" spaced-repetition algorithm?
There's no single algorithm proven best for every person and subject. SM-2 is used here because Woźniak published its exact formula in full in 1990, so it can be reproduced precisely rather than approximated. SuperMemo's own software has used more elaborate algorithms since the 1990s, and other tools such as Anki use their own modified versions of SM-2 — this page implements the original, fully documented version.
What if I don't know my item's current ease factor?
Leave it at the default of 2.50. That is the starting value Woźniak's algorithm assigns to every new item before any repetitions.
Why does a low grade restart the interval?
Per Woźniak's original step 6, a quality rating below 3 means the item wasn't really recalled, so repetitions for it restart from the beginning (interval back to 1 day) as if it were being memorized for the first time. The ease factor itself is only adjusted by that grade's own formula, not reset.
Does this tool save or send my data anywhere?
No. Everything on this page runs in your browser tab only. Nothing is saved to a server, an account, or your device's storage — refresh the page and the schedule is gone. Copy or screenshot it if you want to keep it.
What if I review something late, after its scheduled date?
Woźniak's published algorithm doesn't include a specific rule for overdue reviews, so this tool doesn't invent one. Grade honestly based on how hard the recall actually was — a late review that was still easy earns a high grade; one that felt shaky earns a lower one, and the formula responds accordingly.
05 — Sources
Where this formula comes from
- Woźniak, P. A. (1990). "Optimization of Learning" (Master's thesis, University of Technology in Poznań); republished as "Application of a Computer to Improve the Results Obtained in Working with the SuperMemo Method." SuperMemo World.supermemo.com · published 11 Aug 1990, page updated 2 Nov 2022 · accessed 26 Jul 2026
- Wikipedia contributors. "SuperMemo." Wikipedia, The Free Encyclopedia — see the "SM-2 algorithm" section for an independent restatement of the same formula.en.wikipedia.org/wiki/SuperMemo · accessed 26 Jul 2026
Method: every interval and ease-factor value on this page — in the interactive tool and in the worked example — is produced by one JavaScript function implementing the seven-step SM-2 algorithm exactly as Woźniak stated it in source 1, including his specific instruction to round fractional intervals up. No interval, date or ease factor anywhere on this page is hand-typed or estimated.
Key terms for this topic
Generated on request by this site’s AI from the topic below — not a hand-written glossary. Useful for orientation; check anything you intend to rely on.