SM-2 explained — the algorithm behind memorization that sticks

If you read a verse today and never revisit it, you will have forgotten about 80% of it by tomorrow. This is the forgetting curve, identified by psychologist Hermann Ebbinghaus in 1885. It is universal and merciless.

The SM-2 algorithm (SuperMemo 2), developed by Piotr Wozniak in the 1980s, does not fight this curve: it leverages it. Here is how.

The core principle: review just before forgetting

Intuitively, we review “when we think about it.” Sometimes too early (the verse is still fresh, review is pointless), sometimes too late (the verse is completely gone, review = re-learning from scratch). Both waste time.

SM-2 computes the optimal moment: just before forgetting begins. At that moment, review requires a small recall effort, which strengthens the memory trace far more effectively than a passive re-read.

The equation (simplified)

For each card, SM-2 tracks two values:

EF (Easiness Factor): how “easy” the card is for you. Starts at 2.5.
Interval I: days until the next review.

After each review, you rate your performance from 0 to 5 (0 = total blank, 5 = perfect recall). The algorithm updates:

If rating >= 3 (success):
  I = I * EF  (interval extended)
  EF = EF + (0.1 - (5 - rating) * (0.08 + (5 - rating) * 0.02))

If rating < 3 (failure):
  I = 1  (restart)
  EF unchanged (no punishment)

In practice, for a well-memorized verse:

First review: 1 day later
Second: 6 days later
Third: ~15 days
Fourth: ~38 days
Fifth: ~100 days
And so on, with intervals expanding as long as you succeed.

Why it works for the Bible

Bible memorization has two properties that make SM-2 particularly effective:

Verses are discrete, bounded units — ideal for flashcard format.
They are reused over time — a verse memorized in January is still useful in December, unlike ephemeral knowledge.

5 to 15 minutes a day suffice because the algorithm presents only the critical cards for that day. It does not waste your time on what you already master.

The honest limits

SM-2 is not magic. Three things it does not do:

It does not understand meaning. If you answer by rote without understanding, the verse stays opaque. Memorization must accompany contextual reading and meditation.
It cannot recover a lost card in one day. If you disappear for 6 months, return is painful (many cards to relearn). Regularity > intensity.
It does not replace prayer. The Word enters the heart by the Spirit, not by the algorithm. SM-2 installs the text in memory; the dwelling in the heart is another work.

Beyond SM-2

More recent algorithms exist (FSRS is the most promising, recently adopted by Anki) and better model individual variation. AgapePlay uses SM-2 for now because its parameters are battle-tested, and FSRS gains are marginal on vocabulary as structured as the Bible.

In 5 years, we may switch to FSRS or its successor. But the principle — review just before forgetting — will remain.