Trying to Reason from Anecdotes

Point: in Kansas City, a police officer appears to have shot another officer during an arrest (presumably by accident), then killed the person being arrested; the police then reported that the arrested/dead person had been the one to shoot the cop, thus justifying the shooting. Only it was caught on video. 

Counterpoint: in Volusia County Florida, police had responded 300 times to a juvenile group home, only to be ambushed by three of the juveniles who had fled, broken into a nearby lake house, found guns, and taken up shooting at the police for fun. The three were 12-14 years old. The police did eventually shoot one of them, not fatally, but only after enduring an hour and a half of gunfire themselves. 

The police in the second case have a very clear idea of what went wrong there: they say the juvenile system in Florida is broken. The activists in the first case have a very clear idea of what went wrong there: they say the policing system is broken. 

Once we as a culture would have said that these were too specific for general lessons; rather than try to alter legislation or training, we'd have referred the matters to prosecutors for individual trials on the exact charges in play. That era has apparently passed, but the new era leaves us scrambling to try to reason from anecdotes. That a bad thing can possibly happen leads to laws and restrictions that may make it harder for good outcomes to happen in many other cases. The emotions attending the bad cases lead to irrational decisions applied to the whole model. This is no way to run a railroad.