Change Details

The Revert Risk models are quite sensitive to user groups and edit count, so they generally won't give high scores to users with many edits/groups on a particular wiki. Unfortunately, global editors like Stewards and Global Sysops don't necessarily have local user groups or high edit counts, so they're liable to be reverted by Automoderator because they'll often look like new editors on projects they infrequently edit. In addition to checking local groups, we should be checking for certain global groups and skipping reverts for users who have them.

The Revert Risk models are quite sensitive to user groups and edit count, so they generally won't give high scores to users with many edits/groups on a particular wiki. Unfortunately, global editors like Stewards and Global Sysops don't necessarily have local user groups or high edit counts, so they're liable to be reverted by Automoderator because they'll often look like new editors on projects they infrequently edit. In addition to checking local rights, we should be checking for certain global rights and skipping reverts for users who have them.

The Revert Risk models are quite sensitive to user groups and edit count, so they generally won't give high scores to users with many edits/groups on a particular wiki. Unfortunately, global editors like Stewards and Global Sysops don't necessarily have local user groups or high edit counts, so they're liable to be reverted by Automoderator because they'll often look like new editors on projects they infrequently edit. In addition to checking local groupsrights, we should be checking for certain global groupsrights and skipping reverts for users who have them.