A lot of the discussion of gun control and public safety comes from multiple victim murder incidents like Virginia Tech and Columbine. But, these incidents are a poor basis for policy.
According to the United States Justice Department, in the United States:
In 2005, of all homicide incidents--
* 4% involved 2 victims
* .6% involved 3 victims
* .1% involved 4 victims
* .05% involved 5 or more victims
There were 16,692 homicide victims that year.
These incidents are without a doubt horrible. But, this doesn't justify turning our nation into an armed fortress. At least half of murders in the United States are committed by someone the victim knows.
A significant number of multiple victim murder incidents are also one off affairs. The killer frequently ends up dead, or is captured on the spot or shortly thereafter and spends the rest of his life incarcerated, whether or not he is convicted or executed.
Put it this way. It is safe to guess that everyone who will commit a mass murder in the next decade is alive today. There are fewer than 100 people alive right now who will commit a mass murder in the next decade in the United States. There are 300,000,000 people alive in the United States right now. Some are seriously disturbed individuals for whom intervention is possible before it is too late. But, some of these incidents provide very little warning.
There are tens of millions of people with felony records. There many millions of people who suffer from mental illnesses which from which previous mass murdered have also suffered. There is considerable overlap between these two populations, because of the way that our criminal justice system has been used as a backup alternative to a mental health system. Yet, even a sieve that focused on these populations alone wouldn't cut it. Virginia Tech and Columbine were both committed by people with no criminal record. Many mass murders are committed by people for whom there is no public history of mental illness. Mass murderers are overwhelmingly relatively young men, but that sieve again, isn't nearly fine enough to isolate the 100 people an early warning system would like to identify.
Mass murders are by and large unpredictable and unpreventable, compared to other kinds of murders.
In contrast, a mere 1% drop in murders involving a single victim would save far more lives than a complete elimination of mass murders, and there are enough predictable patterns to some other kinds of murders that reducing them by a small percentage is a much more viable proposition. It is far easier to identify violent gang members or people likely to kill in the context of abusive domestic relationships than it is to identify someone who will errupt into becoming a mass murderer.
Similarly, the prospects of preventing traumatic deaths through improved suicide prevent systems that cut suicides 1-2%, would be much easier.
Likewise, increased public awareness of safety measures directed at accidental deaths like deaths from lightening strikes and floods, that significantly reduce particular kinds of preventable accidental deaths, would be more effective. You'd save more lives by modestly cracking down on reckless teen driving, or drunk driving, or increasing use of trigger locks and locked gun cabinets, than you could by an all out, resource intense effort to prevent mass murders.