I'll tell you what scares me: I don't think we know how to prevent a tragedy like the Newtown massacre. The more information that emerges on the killings, the less effective any of the potential policy remedies appears to be.

The private-sales loophole (which is often called the "gun-show loophole" but is much broader than that) should be closed, but Adam Lanza didn't buy his guns at a gun show. He took them from his mother.

Background checks should be stricter, but again, Lanza would never have been stopped by a background check. The guns weren't registered in his name.

Waiting periods make sense for gun buyers, but again, they would've had little effect on Lanza, who didn't buy the guns he used.

The assault-weapons ban should never have been permitted to lapse, but if it were still law, there's little evidence that it would have stopped Lanza, or even forced him to use a different kind of gun. The ban was full of loopholes, and any gun built before it took effect was grandfathered in as fully legal.

Taking large-capacity magazines off the market - another feature of the assault-weapons ban - would certainly be a good idea, but there's not much evidence that it would help, at least not in the near term. Last time we tried it, manufacturers flooded the market before the ban went into effect, making it unusually easy for anyone who wanted such a magazine to get one.

Some gun enthusiasts have argued that if more people carried guns, Lanza would've been swiftly stopped. Of course, Lanza's mother loved guns and was highly trained in their use. Those guns didn't save her life; they led to her death.

The country could surely do a better job providing mental-health resources and destigmatizing treatment, but there's no evidence, at least not yet, that an inability or unwillingness to get mental-health treatment was a problem for the Lanza family. And we need to be very careful that we don't tip into profiling the mentally ill, who are vastly more likely to be the victims of violent crime than the perpetrators.

Perhaps the wisest single sentence I've read in the aftermath of the shooting came from Mark Kleiman, a crime specialist at the University of California at Los Angeles. "Figuring out how to prevent the next gun massacre (or specifically the next gun massacre at a school) is a classic case of solving the wrong problem," he wrote. "The right problem is gun homicide generally, or homicide generally."

It's a depressing thought, but perhaps it shouldn't be. Yes, in the aftermath of the massacre in Newtown, we want to stop the next one. It is possible, and perhaps even likely, that we can't. Although these rampages have come with searing frequency over the past two years, they shock us so deeply because they remain so rare.

But gun violence isn't so rare. According to the Brady Campaign, in the United States, more than 12,000 people die in gun homicides every year. More than 18,000 kill themselves with guns. Almost 600 are killed in gun accidents. More than 66,000 are injured by guns. These traumas sadden, but they're so common that they no longer shock.

They get, however, to the truth of this issue: While we may not be able to stop every gun death, there are lots and lots and lots of gun deaths to stop. And if a massacre like the one in Newtown is specific and idiosyncratic in ways that make it very difficult to confront through policy, the average gun death follows a much clearer pattern.

The fatalism about gun control tends to begin with a simple statistic: There are 300 million or so guns in the United States. Perhaps it would be better to live in a world where that number was much closer to zero. But since we don't live in that world, the thinking goes, there's nothing much that can be done.

Kleiman doesn't buy that fatalism. Of those guns, 100 million are handguns, and handguns are used in the bulk of killings (though not in the Newtown massacre).

Moreover, Kleiman says, the evidence suggests that these old guns aren't huge contributors to gun crime. "The fact that we have all these guns in inventory doesn't seem to matter much because crime guns are young," he says. "Bad guys like new toys fresh out of the box. Now, maybe they'd adapt if you made those guns hard to get. But your local branch of the Crips isn't arming itself out of the proceeds of burglaries. They're buying new Glocks."

That's where the private-sales loophole comes in. It's depressingly easy for a gang member to drive to a gun show outside the city limits and bring back dozens of Glocks with few questions asked. That's something we can, and should, stop.

As for the kind of guns you can buy, a tougher assault-weapons ban, with fewer loopholes, and perhaps provisions outlawing bullets built to shatter in the body in order to cause maximum damage, would help reduce the lethality of the arms on the street.

Instituting a real waiting period has been successful in reducing suicides. The evidence shows that if people can't get a gun quickly, they often don't kill themselves. A week should do it. And for the life of me, I can't understand why I should need to take lessons and pass a test to drive a car, but I can, in many states, get a gun immediately and with no demonstrated ability to use it safely. So perhaps mandatory training could be part of that package, too.

But even if we do all this, and more, we may still see rampage shootings, and we will still grieve for murdered children. But the shootings will be fewer, and the deaths rarer, and that's all policy can do.

Ezra Klein is a columnist at The Washington Post.