No one can discount the threat of nuclear war, even when the threat is still materializing and remains far lower than some media reports would indicate. At the same time, it makes little sense to define the nuclear threat simply in terms of the range of ballistic missiles and the reliability of their reentry vehicles without considering all of the factors that shape the real-world effectiveness of such weapons, and why and how they might actually be used. Worse, it makes even less sense to focus on nuclear exchanges without regard to the broader strategic context in which such exchanges might take place.
First, there is a very real difference between having a potential or token nuclear-armed missile threat and having a real one. Simply demonstrating a large booster is very different from having a nuclear-armed missile that can actually be used in war. A fully functional missile requires a re-entry vehicle with a warhead that minimizes weight but must handle the movement of the missile in the field, the vibration of the boost phase, and the heat and shock of reentry. Hard decisions have to be taken about arming the warhead, which is most reliable in the prelaunch phase, but can be a nightmare that strikes on your own soil if the missile malfunctions.
Getting a predictable yield and set of effects requires a great deal of simulation at a minimum, and probably the test and recovery of simulated nuclear warhead designs to show the actual effect of firing a missile with such warheads, and the impact of launch and recovery. Plus, it requires telemetry tests of warhead detonation in other missiles to confirm the ability to control the height of burst — which is critical in determining the real-world effect of a nuclear warhead in terms of blast, radiation, thermal, and fall out.
Missile design has advanced a great deal since the first nuclear-armed missiles, but reliability is still a critical issue. A failed launch or flight by a nuclear-armed missile raises major problems, even if the warhead does not detonate. This is particularly true when a nation has only limited nuclear forces. In such cases, the nation has to make hard decisions about how many nuclear missiles to launch against a given target, whether to launch them in sequence or at the same time, and what happens if the missiles malfunction at the desired point of detonation (or hit and explode in very different locations from the ones they were aimed at).
There are a wide range of things that can go wrong that a few “white suit” firing tests from a test range won’t reveal. Missile transportation, aging, and storage effects can all alter missile behavior. Problems that arise in programming the missile launch, and in setting up its Transporter-Erector-Launcher (TEL), seem less likely to be severe, but can occur. Real world missile accuracy requires extensive tests — not simply a few demonstrations. Missile component failure is a constant problem.