Spaced Repetition Learning: Building LearnLoom with Realm DB and Expo

Students who use spaced repetition retain 90% of material vs 20% with traditional study. Here is how we implemented SM-2 in a mobile app.

1. The SM-2 Algorithm

Each card has an ease factor and interval. After each review, the algorithm adjusts when the card should appear next:

typescript
function calculateNextReview(card: Card, quality: number) {
  let { easeFactor, interval, repetitions } = card;
  if (quality >= 3) {
    interval = repetitions === 0 ? 1 : repetitions === 1 ? 6 : Math.round(interval * easeFactor);
    repetitions++;
  } else {
    repetitions = 0; interval = 1;
  }
  easeFactor = Math.max(1.3, easeFactor + (0.1 - (5 - quality) * (0.08 + (5 - quality) * 0.02)));
  return { ...card, easeFactor, interval, repetitions, nextReview: addDays(new Date(), interval) };
}

2. Why Realm DB

Realm DB offers zero-copy reads, meaning flashcard data loads instantly without serialization overhead — critical for decks with 10,000+ cards.

3. Rich Card Content

Cards support markdown, LaTeX formulas, and embedded images, making them suitable for medical and engineering students.

Summary

Combining proven learning science with fast local databases creates a study app that genuinely improves long-term knowledge retention.