The U.S. can prevent a North Korean nuclear attack either by deterring one, or by eliminating the North Koreans' ability to launch one through military means. The foreign policy establishment's view is that because deterrence worked on the Russians it will also work on the North Koreans, making deterrence the obvious choice.

Yet given how deterrence with the Russians actually worked, it is doubtful that deterrence of North Korea is possible. It follows that military action will be required to avert or at least minimize the consequences of an inevitable nuclear war, and better sooner than later.

Even assuming that Kim Jong Un is fully rational, and that the logic of mutually assured destruction can prevent him from launching a nuclear attack on purpose, successful deterrence requires far more than that.

The U.S. and Russian experience shows that technical malfunctions, human error, and military misunderstandings will inevitably generate false alarms that an enemy attack is underway. The U.S. and Russia have avoided nuclear war (so far) only because each has developed a process that lets them vet apparent threats and identify false alarms before launching (what each would think is) a retaliatory strike. The false alarms and nuclear close calls have been many.

What's more, the vetting of these alarms requires time and military officers with an independence and a professional ethos that enables them to refuse to act upon an alarm that they believe is false no matter what official doctrine requires.

For example, in 1983, the officer in charge at the Soviet early warning center (Stanislav Petrov) received a satellite alarm indicating that the U.S. had launched a first strike. The satellite appeared to be working correctly. In this case, it was Petrov's duty to pass the warning on to his superiors, who were in turn obligated by Soviet doctrine to launch a nuclear counter-strike. Instead, Petrov independently decided that the alarm just had to be false. Absent either the time to think the situation through or a military ethos by which Petrov could exercise his independent judgement, it is highly likely that this incident would have led to the Soviets launching a first strike on the U.S.

To put this another way, deterrence "worked" against the rational Russians only because we were also enormously lucky.

So, can North Korea be deterred?

North Korea is a very small country on the ocean adjacent to the U.S. Navy and U.S. allies. These brute geographic facts mean that the North Korean officers in the nuclear command chain will need to respond almost instantly to any alarm they receive. The Kim regime perceives its biggest threat to be internal subversion, and so the very last thing that this regime will do is to train military officers to act independently upon their own judgement. In short, North Korea is structurally incapable of evolving a functional alarm vetting process.

North Korea's inability to reliably vet alarms, and a policy of confrontation that will surely produce false alarms in abundance, implies that a North Korean nuclear first strike on the U.S. is statistically inevitable — no matter how lucky we are.

A war with North Korea is coming either way.

And so the U.S. has two broad policy options to deal with this reality: It can respond after a North Korean nuclear strike triggered by a false alarm, or it can launch a pre-emptive strike that eradicates the Kim regime and degrades North Korea's ability to inflict harm upon the U.S. and its allies to the greatest extent possible.

Given North Korean defensive measures, this effort will almost certainly require the use of tactical nuclear weapons. As horrible as a war with North Korea will be, the war will undoubtedly be far less horrible if it begins with a U.S. pre-emptive strike rather than with a North Korean first strike. It is as simple as that.

The devastation that a North Korean first strike would potentially inflict becomes greater every day. And the extent to which a U.S. pre-emptive strike can degrade the North Koreans' ability to inflict harm becomes smaller each day. It follows that the best option the U.S. has to deal with the North Korean threat is to nuke North Korea now.

Kevin R. James is a Research Fellow in the Systemic Risk Centre at the London School of Economics.

If you would like to write an op-ed for the Washington Examiner, please read our guidelines on submissions.