Spaced Repetition: The Algorithm That Beats the Forgetting Curve

Here’s an uncomfortable fact about your brain: within 24 hours of learning something new, you’ll forget most of it. Within a week, without review, the majority of it is gone. This isn’t a personal failing — it’s the forgetting curve, one of the oldest and most replicated findings in psychology.

The good news is that the forgetting curve has a well-documented weakness, and exploiting it is the single highest-leverage change you can make to how you study.

The forgetting curve, briefly

In the 1880s, German psychologist Hermann Ebbinghaus spent years memorizing lists of nonsense syllables and testing himself at intervals. What he found — and what modern replications have confirmed — is that memory decays predictably: fast at first, then more slowly. Freshly learned material bleeds away within hours; the little that survives a week tends to stick around much longer.

That second part is the key. Each time you successfully recall something just before you would have forgotten it, the memory becomes more durable. The forgetting curve flattens. What would have lasted a day now lasts a week; what lasted a week now lasts a month. Memory researchers call this the spacing effect, and a meta-analysis by Cepeda and colleagues (2006), covering 254 studies and more than 14,000 participants, found spaced practice reliably beats massed practice — cramming — across nearly every kind of material tested.

Why cramming feels good but fails

Cramming works for exactly one thing: tomorrow morning’s exam. Massed repetition inflates your current recall while doing very little for durability — the memory equivalent of a sugar high. This is also why rereading your notes feels productive: the material becomes fluent and familiar, and your brain mistakes that fluency for knowledge. Psychologists call this the illusion of competence. The material is available now, so you assume it will be available later. The forgetting curve says otherwise.

Durable memory is built from a different ingredient: effortful recall at increasing intervals. Struggling to retrieve something — and succeeding — is what signals your brain that this memory is worth keeping.

From flashcards in shoeboxes to algorithms

Learners have exploited the spacing effect for decades. In the 1970s, Sebastian Leitner formalized it with a shoebox: flashcards move through compartments, and each compartment is reviewed less often than the last. Answer correctly, the card moves back a compartment; miss it, and it returns to the front.

Software made the schedule per-card and precise. The SM-2 algorithm (1987), which powered early versions of Anki and SuperMemo, assigned each card an “ease factor” and stretched its review interval with every success: one day, six days, then multiplying outward.

The state of the art today is FSRS (Free Spaced Repetition Scheduler), the memory model behind modern Anki — and the one Quizatto uses. Instead of a fixed multiplication rule, FSRS models each card with two quantities:

• Stability — how many days until your recall probability decays to 90%
• Difficulty — how hard this particular item is for you, on a 1–10 scale

Its parameters were fit on hundreds of millions of real reviews from real learners. Every time you rate a card — Again, Hard, Good, Easy — the model updates both numbers and schedules the next review for the moment your recall probability is predicted to dip below the target. Not sooner (wasted effort), not later (you’d have forgotten).

What this looks like in practice

Say you learn the Spanish word la manzana (apple) today.

1. Today: you see the card, recall it, rate it Good. The algorithm schedules it for ~3 days out.
2. Day 3: you recall it again with a little effort. Next review: ~10 days.
3. Day 13: still there. Next review: a month.
4. After a few more cycles, the interval is measured in months — and that word is functionally yours for life, at a total cost of maybe two minutes of review time.

Multiply that across a thousand cards and you get the real promise of spaced repetition: near-permanent retention for minutes a day, because on any given day only the cards that are actually due — the ones at the edge of forgetting — show up.

How to use it well

• Review daily, briefly. Ten focused minutes daily beats an hour on Sunday — the schedule only works if you show up when cards are due.
• Rate honestly. If you hesitated, that’s Hard, not Good. The algorithm can only model your memory from what you tell it.
• Don’t fear the “Again” button. Failing a card isn’t a setback; it’s exactly the information the scheduler needs.
• Combine it with retrieval practice — spaced repetition is retrieval practice with perfect timing.

The forgetting curve is not negotiable. But its shape is — and the tools to reshape it now fit in your pocket. Quizatto runs FSRS under every deck you study, on iPhone and the web, so the only thing left to bring is the ten minutes.