• 8 Posts
  • 18 Comments
Joined 2 years ago
cake
Cake day: September 23rd, 2023

help-circle


    1. Compute raw post score (upvotes minus downvotes).
    2. Normalize score by community size (e.g., divide by square root of active users).
    3. Calculate z-score relative to community mean and standard deviation.
    4. Apply time decay to prioritize recent posts.
    5. Sort posts by adjusted z-score.
    • Outcome: Posts that significantly outperform their community norm appear prominently, giving small and large communities equal visibility potential.
    • Enhancements: Minimum engagement thresholds, Bayesian shrinkage for small communities.