The Corner

The one and only.

Someone’s Been Watching Too Much Numb3rs


From The Atlantic, by a man who’s done a lot of work with marketing data:


Big data might have stopped the massacres in Newtown, Aurora, and Oak Creek. But it didn’t, because there is no national database of gun owners, and no national record-keeping of firearm and ammunition purchases. Most states don’t even require a license to buy or keep a gun.

That’s a tragedy, because combining simple math and the power of crowds could give us the tools we need to red flag potential killers even without new restrictions on the guns anyone can buy. Privacy advocates may hate the idea, but an open national database of ammunition and gun purchases may be what America needs if we’re ever going to get our mass shooting problem under control.

Just look at the gun-acquiring backgrounds of some of our more recent mass killers to see what I mean. James Holmes, the Aurora shooting suspect, went to three different locations spread out over 30 miles to legally buy his four weapons. All three were reputable outdoors retail chain stores. He then went online, and bought thousands of rounds of ammunition along with assault gear. . . . In Newtown, Adam Lanza carried hundreds of rounds — enough to kill every student in the Sandy Hook Elementary school if he had not been stopped. . . .


Let me give you some numbers. According to the ATF, about 4.5 million new firearms, including approximately 2 million handguns, are sold in the United States every year, along with roughly 2 million more secondhand ones. No one knows how much ammunition is sold yearly in the country, but as a yardstick, the U.S. Military bought 1.8 billion rounds in 2005.

That sounds like a lot — but it isn’t. Not in today’s data-driven private sector. This data would fit on a thumb drive you can buy at any Best Buy and still leave room for the transcripts of every Rush Limbaugh show ever.

I don’t think anyone doubts that modern computers are capable of handling a lot of data. But handling data and collecting data are two different things. It would take a lot of work to enter every single ammunition and firearm purchase into a database, and it would be exceedingly difficult to ensure accuracy. Transfers of guns and ammo between private citizens would be even harder to keep track of.


Analyzing the data is yet another problem. We’d need to decide how much firepower is “too much” and what to do about people who exceed the limit (especially without actual “new restrictions on the guns anyone can buy”). An enthusiastic shooter (or a family of them) might go through “hundreds of rounds” in a single session at the range; any system that flagged Adam Lanza (or his mother) would also generate a ton of false positives.

It’s a little more plausible that such a database could pick up on massive ammo stockpiling by someone like James Holmes — but one has to ask how many James Holmeses there are (could the funding for this project save more lives if put to a different use?) and whether they would find other sources, or make do with less ammo, if they knew gun stores were tracking their purchases.

Similar problems plague the meth-precursor database, which the author mentions as an example of how such projects can work. When a meth cook can’t buy the Sudafed he wants, he sets up a purchasing network, turns to the black market, or finds a substitute; if that fails, meth cartels from other countries pick up the slack. Further, meth cooking is far more common than mass shootings, it requires large quantities of pseudoephedrine, and the Sudafed crackdown comprises not just purchase tracking but also purchase limits low enough to ensnare legitimate users. Even the limited and debatable success of this effort — it might have reduced domestic meth production, though probably not consumption — would be hard to achieve in the realm of mass shootings, which are incredibly rare and do not require an ammo stash big enough to stand out in a database.

It is very difficult to control illegal substances that are in high demand, to say nothing of common, legal products with legitimate uses. The magic of “big data” doesn’t change that.



Sign up for free NRO e-mails today:

Subscribe to National Review