We have a new policy of troll discouragement, and as one might expect in new situations, the standards appear to be fuzzy. Perhaps concrete examples might help. The following is a very common kind of trolling response: Note the features: the troll discovers "implications" not overtly visible in someone's post, and the discovered implications are of a specific kind: they are an imputation of deliberate, intentional evil, evil as a goal; which the troll then declares ridiculous, and uses to attack the poster for harboring such ideas. Common encounters: Someone says the US is killing too many civilians in Iraq, shooting too many pregnant women at checkpoints, etc. The troll replies that accusing the US of intentionally killing civilians such as pregnant women in Iraq is ridiculous, demands evidence of such intent. Someone says too many black men are being jailed, and it is wrecking he black communities n the US. The troll replies that accusing the police departments in the US of intentionally destroying local black communities is ridiculous. And so forth. This is one of the most common rhetorical means of deflecting a thread into inflamed and off topic argument. Perhaps we could use it as a touchstone to guide policy toward such deflections?