Puzzle

How we calculate. Each source is normalized to a 0–100 scale, then we take a weighted average. Critic-only sources get a slight discount when they disagree dramatically with audience scores. Sources with fewer than 25 ratings are excluded.