Robert Gard and Philip Coyle label the U.S. Ground-Based Midcourse Defense (GMD) system a “mistake,” despite last month’s successful intercept of a target missile. The system has not infrequently failed tests in the past, you see.

And so they appear to argue: When the system fails a test, it proves it is not good enough. And when the system succeeds, it proves that the program is in too much of “a rush” and not “taking the time to get it right.”

It’s not a novel argument, and far from convincing. The latest GMD test was not only successful, it was also the most challenging test to date—proof that the program is, in fact, making significant progress toward “getting it right.” The GMD interceptor defeated countermeasures during this test. This is an extremely challenging task, due to all the background noise and debris in space where the intercepts occur.

While criticizing the system’s development history as a “rush to failure,” Gard and Coyle omit a critical fact. Deployment of the system was fast-tracked because the United States was critically vulnerable to North Korean, and potentially Iranian, long-range ballistic missiles. Both countries had demonstrated capabilities that would allow them to develop long-range ballistic missiles. The only real alternative to standing up to a less-than-perfect missile-defense system is to cede your adversaries the ability to obliterate millions of people within minutes. That is clearly not a good option.

Despite tremendous international efforts, North Korea has not gotten any less aggressive since the 2002 decision to deploy a U.S. missile-defense system. Pyongyang has tested nuclear devices in 2006, 2009 and 2013. It has also extended the reach of its long-range ballistic missiles through its ongoing testing regimen.

As for terror-sponsoring Tehran, Gard and Coyle state that “the State Department is working to reduce the scope of Iran’s nuclear program to clearly peaceful civil purposes.” The United States should not entrust millions of American lives to the State Department’s diplomatic prowess. After all, the State Department also worked to make North Korea’s nuclear program for peaceful civil purposes. It worked to reduce the scope of North Korea’s ballistic-missile program. Those efforts have proved futile.

In 2013, the Defense Intelligence Agency assessed “with moderate confidence the North currently has nuclear weapons capable of delivery by ballistic missiles.” No wonder the Bush and the Obama administrations agree on the criticality of missile-defense systems. No wonder President Obama’s Pentagon has doubled down on the GMD system. After all, it is currently the only missile-defense system capable of dealing with long-range ballistic missiles. And since the United States spends less than 2 percent of the Pentagon’s dwindling budget on missile defense, the situation is not likely to change anytime soon.

Indeed, the United States needs a more robust testing program. During ballistic-missile tests, we need to push the performance envelope of the system. This means that the tests are more likely to fail. And it’s ok. Design a test so that you can’t fail, and you’re bound to “win,” but you’ll also learn nothing (and waste millions of dollars in process). The ultimate metrics of a successful test should be whether and how much we have learned.

The Polaris AX submarine-launched ballistic missile (SLBM) had twelve failures within a span of a year. Yet, we learned from those failures and perfected the system to the point where many of its technologies were used in the Trident SLBM. Is Trident a failure because it is based on a missile that failed 70 percent of the time? Hardly; it is expected to remain in service until 2040.

Bad countries with ballistic missiles are not going away. As Vice Admiral James D. Syring, director of the Missile Defense Agency, succinctly wrote: “the [ballistic missile] threat continues to grow as our potential adversaries are acquiring a greater number of ballistic missiles, increasing their range and making them more complex, survivable, reliable, and accurate.”

Failing a test is not catastrophic; it is how science—and security—advances. The GMD system has already done what many said was impossible: identified, tracked and killed an enemy warhead in flight—a dart hitting a dart.

Yes, we must continue to improve and perfect the system, but we also must continue to deploy the capabilities we do have that can keep us one step ahead of the threat. To do otherwise would be to leave ourselves vulnerable to a ballistic-missile attack—inviting a catastrophe too grave to contemplate.

 - Michaela Dodge is an analyst specializing in defense and strategic policy issues in The Heritage Foundation’s Douglas and Sarah Allison Center for Foreign and National Security Policy.