MeatballWiki | RecentChanges | Random Page | Indices | Categories

A person who has been in constant conflict with the community suddenly seems to repent. She apologizes, stops violating CommunityExpectations and starts making constructive contributions. Are we witnessing a reformed troll?

It's certainly possible. However, it may also be the TrollingTactic of FalseRepentance. By ceasing all disruptive behaviour, a troll can sow confusion in the community. In most cases, there will be a reaction spectrum. On one end are the people who want to ForgiveAndForget and AssumeGoodFaith. They will welcome the seemingly reformed troll with open arms. On the other end are the skeptics. They refuse to accept this turnaround, and perhaps even continue to be openly hostile. The rest of the community falls somewhere in the middle: happy to see the end of the disruption, but wary of the former troublemaker. The advantages for the troll provided by this tactic vary according to how the community is distributed on the spectrum.

Particular tactics

Beware that the troll has the attacker's advantage of flexibility - she can start off being angelically good, then move to betrayal or alienation later on, according to what seems likely to be most effective. FalseRepentance can also be finnessed into an ExitStrategy, as a precursor for leaving a community and dodging negative consequences.

If the distribution is quite even, the troll may decide to be angelically good, apparently striving to become a model CommunityMember. This creates significant tension between the people on opposite ends of the spectrum; the skeptics are sickened at how others could be taken in by the performance, while the forgivers question the motives behind the continued hostility of the other members. The people in the middle may find themselves quietly taking sides, and the issue becomes a LandMine. At this point, the troll can exploit the tension and encourage conflict by subtly emplying other TrollingTactics, such as SelectiveFlattery or the SympathyPlay?.

If more people fall on the ForgiveAndForget side, the troll can use the weapon of betrayal. If, after repenting, the troll goes back to his anti-social ways, the community will likely feel hurt and angry, making them more likely to respond to further provocation; few will be able to DissuadeInteraction because flaming back is more acceptable to a betrayed person's sense of vengance.

With more people distrustful of the apparent repentance, the troll may try to encourage the alienation of the few that want to AssumeGoodFaith. As most of the community is skeptical, those who accept that the troll has reformed may find themselves outcasts, at least temporarily. The troll can attempt to become their ally, creating resentment in those few against the rest of the community. Even after the troll's plan becomes obvious, this rift in the community may be difficult to heal.


How can we tell if a difficult person has truly had a change of heart, or is just using the FalseRepentance tactic? We don't want to drive away someone who is truly repentant, nor do we want our GoodFaith exploited. Some things to keep in mind:

FalseRepentance is one tactic in the overall strategy of the HighLowPump?.


I don't get it. Is this truly a common pattern? Is there any way we can ever hope to differentiate motives? To assume the repentance is false is to ascribe a motive. Few people, few trolls, are complex enough or sufficiently much in control of their personality to carry on a false repentance for any length of time. How can we be sure that the trolling is a genuine trait and the repentance an artificial one? Many posters and participants occasionally use TrollingTechniques? despite a genuine desire to contribute.

That's very simple. You look for repeating patterns. If someone does it once or twice, fine. If he does it 5+ times in 3+ communities, then he follows a pattern. Trolls often forget that their behaviour is observed and remembered. -- HelmutLeitner



MeatballWiki | RecentChanges | Random Page | Indices | Categories
Edit text of this page | View other revisions