X does not confirm shadowbans, so every answer here is inferred from visibility rather than read from an API. Four checks run against four independent data sources: whether the account’s posts appear in search, whether views on its own timeline fall below what its follower count implies, whether its replies are being buried relative to its posts, and whether the handle appears in search suggestions. Two of the four are threshold tests, and this page prints the thresholds.
The ghost-ban test compares average views against a deliberately low floor: an account under 10,000 followers is expected to reach at least 4% of them per post, falling to 0.3% above 5 million followers, because reach per follower drops as an audience grows. Averaging under 10% of that floor is reported as a likely ghost ban and under 30% as a possible one. Accounts below 500 followers are passed without a verdict, because there is not enough traffic to read.
Reply deboosting is a ratio, not a level. Replies normally draw roughly 10 to 30 percent of the views a standalone post gets, so the check fails only below 2% and warns below 5%. Those bands are our model’s parameters, chosen to be hard to trip by accident, not numbers X has published.
Context for the follower thresholds above: across the 18,486,455 X accounts PlayerSells indexes, measured on 24 August 2026 over a 45,308-account sample, the median account has 550 followers and 7,085 is the 90th percentile. The overwhelming majority of accounts therefore sit in the first band of the table below, where the expected floor is highest and a real suppression shows up most clearly.
View floors and ratio bands are parameters of the PlayerSells detection model, not statements published by X. Follower distribution figures are our own measurement over the PlayerSells X index, last computed on 24 August 2026.