The algorithm of X has been made public. Here, I am sharing a summary of the analysis by AlexFinnX in Korean. I will save it with @irys_xyz and refer back to it repeatedly. Key Concepts 🔸The higher the score, the greater the reach; the lower the score, the more limited the reach. 🔸New accounts start at -128 points, and must reach at least 17 points to be visible. Factors Reducing Reach 🔸Use of unpleasant text, topics, or words → up to 80% reduction in reach. 🔸Unpleasant usernames, tweets in all caps → significant penalties. 🔸Including links → penalties. 🔸Low-quality text (typos, etc.), spam, comments from accounts not followed. 🔸Account labeling: low quality, spam, toxic, NSFW, horrific content, repeated mentions of specific individuals, etc. 🔸User reactions: reports, blocks, mutes, marking "not interested/see less" → 3 months of reduced reach. Factors Increasing Reach 🔸Videos (significant boost if watched for more than 10 seconds). 🔸Increased visibility if text posts are read for more than 2 seconds. 🔸Quotes and bookmarks provide a strong boost. 🔸Accounts with blue checkmarks and legacy verified accounts are prioritized. 🔸If recognized as a "viral account," additional boosts are given. Behavior Guidelines 🔸Prohibition of unpleasant language. 🔸Maintain and build the tweepcred reputation score. 🔸Encourage good engagement centered around quotes and bookmarks. 🔸Focus on quality and time spent (over 10 seconds) rather than quantity for media. 🔸Maintain trust to avoid being blocked or muted. 🔸Use accounts with blue checkmarks. 🔸Create content that can hold attention for more than 2 seconds. In summary, the X algorithm significantly varies reach based on **reputation score (tweepcred)** and content reaction metrics, and the penalty system (labeling, negative user reactions) still exists. It has been somewhat proven that including links in posts leads to debusting, which was not clearly understood before.