Skip to content
TheChatArchiveThe Chat Archive
grok-3-preview-02-24· xAIHugging Face LMArena Human Preference 140K

what spaced repetition algorithm is used by Memrise?

From a Chatbot Arena battleThe same person sent this prompt to two anonymous models at once. Compare it with Mistralmistral-medium-2505’s answer.
  1. User

    what spaced repetition algorithm is used by Memrise?

  2. Assistant

    Memrise uses a spaced repetition algorithm that is based on principles similar to those found in other spaced repetition systems (like Anki or SuperMemo), but it has customized its approach to fit its platform and user experience. While Memrise does not publicly disclose the exact details of its algorithm, we can infer some key aspects based on its functionality and general spaced repetition research.

    Key Features of Memrise's Spaced Repetition Algorithm:

    1. Spaced Intervals:

      • Memrise schedules reviews of learned material at increasing intervals, following the core principle of spaced repetition. Items are reviewed just at the point where you are likely to forget them, optimizing retention.
      • The intervals are likely determined by a combination of user performance and a pre-set formula, though the exact formula is proprietary.
    2. Performance-Based Adjustments:

      • The algorithm adjusts the timing of reviews based on how well you perform on each item. If you recall an item correctly and quickly, the interval before the next review increases. If you struggle or forget, the interval shortens, and the item is prioritized for more frequent review.
      • This is similar to the "ease factor" concept in algorithms like SuperMemo's SM-2, though Memrise may use a modified or entirely custom version.
    3. ** Leitner System Influence**:

      • Memrise's system has similarities to the Leitner system, where items are metaphorically moved between "boxes" of increasing difficulty. Correctly recalled items move to a box with longer review intervals, while incorrect items are demoted to a box with shorter intervals.
    4. Gamification and Engagement:

      • Unlike more rigid spaced repetition systems, Memrise incorporates gamification elements (e.g., points, streaks, leaderboards) to keep users engaged. This may influence how the algorithm prioritizes certain items or adjusts intervals to maintain motivation, potentially deviating from purely scientific spacing models.
    5. "Watering" Metaphor:

      • Memrise uses a plant-watering metaphor, where items are "planted" in your memory and need to be "watered" (reviewed) at specific intervals to grow. This is a user-friendly representation of spaced repetition, and the algorithm likely uses this framework to determine when items need review.
    6. Forgetting Curve:

      • Like other spaced repetition systems, Memrise's algorithm is likely based on the Ebbinghaus forgetting curve, which models how quickly humans forget information over time. The goal is to schedule reviews at the optimal point to reinforce memory before forgetting occurs.

    Comparison to Other Algorithms:

    Memrise's algorithm is likely inspired by, but not identical to, well-documented spaced repetition algorithms like SuperMemo's SM-2 (used by Anki) or later SM algorithms. Here are some differences and similarities:

    • Similarities to SM-2:

      • Uses performance-based interval adjustments.
      • Increases intervals exponentially for correctly recalled items.
      • Reduces intervals for forgotten or poorly recalled items.
    • Differences from SM-2:

      • Memrise prioritizes user engagement and may adjust intervals to encourage daily use, even if it deviates from optimal memory science.
      • The exact formula for interval calculation and "ease factor" adjustments is not publicly available, unlike SM-2, which is open-source and well-documented.
      • Memrise may incorporate additional factors, such as the type of content (e.g., vocabulary, grammar, or multimedia) or user behavior (e.g., streaks, session length), into its scheduling.

    Limitations and Customizations:

    While Memrise's algorithm is effective for many users, it is not as customizable or transparent as systems like Anki, where users can tweak parameters like interval growth rates, ease factors, and review caps. Memrise's algorithm is designed to be user-friendly and accessible, which may come at the cost of flexibility for advanced users.

    Conclusion:

    Memrise uses a proprietary spaced repetition algorithm that is inspired by established models like SM-2 and the Leitner system but is tailored to its gamified, user-friendly platform. It emphasizes performance-based interval adjustments, engagement, and accessibility, though the exact details of its formula remain undisclosed. If you're looking for a more customizable or scientifically rigorous spaced repetition experience, you might consider alternatives like Anki, but Memrise is well-suited for casual learners and those motivated by gamification.

Responses were produced in different, uncontrolled contexts (system prompts, settings, prior turns). This is not a benchmark.