The forgetting curve: where it all started
In 1885 the German psychologist Hermann Ebbinghaus published “On Memory”. He memorised lists of nonsense syllables and tested how many he still remembered after different delays. The result was the forgetting curve: most information is lost in the first hours and days after learning, and then forgetting slows down.
His second finding matters even more: every review makes the curve flatter. After a review, material is forgotten more slowly than after the first encounter. So the goal is not to review more, but to review at the right moments.
The spacing effect: what the research shows
Spreading reviews out over time leads to stronger memory than repeating the same number of times in a row. This is called the spacing effect, and hundreds of experiments confirm it.
- Cepeda and colleagues, 2006. A meta-analysis of 317 experiments from 184 articles on remembering words and facts. Distributed practice consistently beat cramming, and the best gap between reviews grew with how long the material had to be remembered.
- Cepeda and colleagues, 2008. Over 1,350 people learned facts, reviewed them after a gap of up to 3.5 months and took a test up to a year later. The best gap was about 20 to 40% of the retention period for a one week test, and about 5 to 10% for a one year test.
- Dunlosky and colleagues, 2013. A review of ten popular study techniques, from rereading to highlighting. Only two received the highest utility rating: distributed practice and practice testing.
Why it helps to almost forget a word
It sounds odd, but reviewing a word you already know perfectly does very little. Psychologists Robert and Elizabeth Bjork describe this with the idea of “desirable difficulties”: memory has storage strength (how deeply a word is learned) and retrieval strength (how easily it comes to mind right now). The harder, yet still successful, a recall is, the more storage strength grows.
So the ideal moment for a review is when the word has started to slip away but you can still pull it back. Review too early and there is no effort. Review too late and you have to learn the word again.
How you review matters too. In a study by Karpicke and Roediger (2008, Science), students learned Swahili word pairs. Rereading after learning did almost nothing for recall a week later, while repeated attempts to recall gave a large boost. Memory is strengthened by retrieval, not by another look at the answer. That is why a card shows the word first and reveals the translation only after you have tried to remember it.
From Leitner boxes to the SM-2 algorithm
In the 1970s the journalist Sebastian Leitner suggested studying flashcards in several boxes: remember a card and it moves to the next box, which you open less often; forget it and it goes back to the first one.
In 1987 the Polish researcher Piotr Woźniak wrote SuperMemo and the SM-2 algorithm, which calculates intervals for every card separately:
- the first review comes after 1 day, the second after 6 days;
- every next interval is the previous one multiplied by the card's “easiness factor”, which starts at 2.5 and never drops below 1.3;
- after each answer you grade how easily you remembered, and the easiness factor goes up or down a little. Hard words come back more often, easy ones less often.
SM-2 still powers many apps, including Anki. Newer algorithms such as FSRS exist, but they all share the same principle: review on a schedule that adapts to your memory.
How it works in NewLexi
NewLexi builds on the idea of SM-2 and keeps the schedule simple and clear:
- Every word has its own schedule. The pause before the next review grows step by step: 1, 3, 7, 15, 30, 60 days.
- Your answer sets the interval. After a card there are four buttons: “Forgot” sends the word back to the first step, “Hard” keeps it on the same step, “Recall” moves it one step up, “Easy” skips a step.
- Four mastery levels (new, learning, almost, known) show how confidently you know each word, so you can see your deck grow.
- A limited daily plan. It includes the words due today and a few new ones, so tomorrow's load does not double. Miss a day and nothing is lost: the words go to the top of the next plan.
- Review in context. The words you learn come back not only on cards but also in stories and dialogues that AI writes at your level. Recalling a word inside a real sentence is one more desirable difficulty.
More about the schedule: how spaced repetition works and four mastery levels of a word.
How to get the most out of reviews
- Recall first, then flip. Even a failed attempt to remember strengthens memory more than peeking at the answer.
- Answer honestly. “Hard” is not a penalty, it is a hint for the algorithm. Overrate your answers and words will come back too late.
- A little every day. 15 minutes daily works better than two hours once a week, because intervals are counted in days.
- Do not fear forgetting. Forgot a word and it went back to step one? That is normal, not a setback: the next recall will make it stick harder.
- Meet words in text. Read stories and go through dialogues: a word you have met in a sentence is easier to use yourself.
Sources
- Ebbinghaus, H. (1885). Über das Gedächtnis. Leipzig: Duncker & Humblot. psychclassics.yorku.ca/Ebbinghaus/index.htm
- Cepeda, N. J., Pashler, H., Vul, E., Wixted, J. T., & Rohrer, D. (2006). Distributed practice in verbal recall tasks: A review and quantitative synthesis. Psychological Bulletin, 132(3), 354–380. doi.org/10.1037/0033-2909.132.3.354
- Cepeda, N. J., Vul, E., Rohrer, D., Wixted, J. T., & Pashler, H. (2008). Spacing effects in learning: A temporal ridgeline of optimal retention. Psychological Science, 19(11), 1095–1102. pubmed.ncbi.nlm.nih.gov/19076480
- Karpicke, J. D., & Roediger, H. L. (2008). The critical importance of retrieval for learning. Science, 319(5865), 966–968. doi.org/10.1126/science.1152408
- Dunlosky, J., Rawson, K. A., Marsh, E. J., Nathan, M. J., & Willingham, D. T. (2013). Improving students’ learning with effective learning techniques. Psychological Science in the Public Interest, 14(1), 4–58. doi.org/10.1177/1529100612453266
- Bjork, R. A., & Bjork, E. L. (1992). A new theory of disuse and an old theory of stimulus fluctuation. In From Learning Processes to Cognitive Processes, Vol. 2, 35–67. bjorklab.psych.ucla.edu/research
- Woźniak, P. (1990). Optimization of learning: Algorithm SM-2. super-memory.com/english/ol/sm2.htm
