Reputation, flags, and trust signals

Reputation is not a popularity contest wearing a fake mustache. It is a factory-scoped trust layer tied to useful engagement, moderation outcomes, and member behavior.

• Positive signals

  Helpful participation fills a reputation bucket and can raise trust over time instead of handing out instant sainthood for one good comment.

  • Positive reactions: useful reactions can give the reactor reputation progress and give the author more when another member reacts.
  • Good reports: when a flag is reviewed and approved, reporters earn points and reputation because cleanup work counts too.
  • Sustained usefulness: accumulated reputation points can raise the score one point at a time.
  • Trusted range: members at 70 or higher can show trusted status; 90 or higher becomes exemplary.
  • Natural drift: after inactivity, scores above or below 80 drift back toward the starting baseline.
• Negative signals

  Bad behavior can lower trust, especially when the moderation queue confirms the room was right to complain.

  • Downvotes: post and comment downvotes are negative reputation events in the model.
  • Thumbs-down reactions: they can reduce reputation progress for the author without taking the bucket below zero.
  • Removed content: removed posts carry a stronger reputation penalty than ordinary negative feedback.
  • Confirmed reports: approved flags can penalize the author and mark the flag outcome as actioned.
  • False reports: malicious flagging can cost the reporter points and reputation because weaponized moderation is still bad behavior.
• Level armor

  Reputation losses are softened for higher-level members because a long contribution history should matter.

  • Levels 1-5 take the full reputation loss.
  • Levels 6-15 take 75 percent of the loss.
  • Levels 16-30 take half of the loss.
  • Levels 31-50 take 25 percent of the loss.
  • Levels 51+ take 10 percent of the loss, which is the Factory saying: this regular has receipts.
• Flag submission

  Members can flag problems without turning every bad post into a public trial with snacks.

  • Reasons include spam or advertising, harassment or bullying, inappropriate content, misinformation, off-topic content, and other.
  • Flags store the reported content, author, reporter, reason, optional details, channel or topic context, and pending status.
  • Logged-in Factory members can flag messages, replies, comments, forum topics, news links, and member issues where those controls exist.
  • Flagging runs through the community action guard, so low reputation, frozen reputation, bans, suspensions, and daily limits can block it.
  • DM reports use a separate safety path so Factory owners see the reported message without getting the entire private conversation.
• Owner review

  Flags land in the moderation queue grouped by content, with reporter details and author reputation visible before the owner makes the call.

  • Ignore flags: dismiss the report and prevent that content from being flagged again.
  • Delete content: approve the flags, delete the content, reward reporters, and penalize the author.
  • Agree with flags: hide the content, reward reporters, and penalize the author without necessarily deleting the record.
  • Ban member: for member reports, owners can agree with the flags and ban the reported member.
  • Disagree with flags: keep the content and penalize malicious flaggers when the report itself is the problem.
• Member impact

  Reputation changes the practical experience inside the Factory, not just the decoration on a profile.

  • Trusted members are easier to recognize when new people enter the room.
  • Uncertain members give owners an early signal before the situation becomes a cleanup project.
  • Below 30 reputation, the member is treated as soft-banned by the action guard.
  • Soft-banned or frozen reputation pauses posting and reporting.
  • Factory owners get a moderation signal that is better than grudges, popularity contests, and whoever typed the loudest paragraph.