Nuclear Strategy
The Logic of Threatening to End the World
The entire peace of the Cold War rested on a doctrine that only works if everyone believes you are genuinely insane enough to use it.
The Idea
Mutually Assured Destruction — MAD, an acronym that was surely chosen with some self-awareness — is often described as a deterrent, but that softens what it actually is: a civilisational hostage arrangement. The premise is deceptively simple. If launching a nuclear strike guarantees your own annihilation in return, no rational actor will ever strike first. Peace emerges not from goodwill but from the mathematics of mutual suicide. What makes this genuinely strange is that MAD only functions if your enemy believes you will retaliate even after you've already been destroyed. This is the problem strategists call 'second-strike credibility.' A retaliatory threat has to be credible after your cities are gone, your leadership is dead, and the rational case for striking back has collapsed — because the damage is already done. To solve this, the US and Soviet Union both invested heavily in making retaliation automatic, dispersed, and survivable: nuclear submarines that couldn't be taken out in a first strike, hardened silos, airborne command posts. The theorist Herman Kahn went further, arguing that you had to make your enemy believe you might be irrational — that you might retaliate even when it's pointless. This became known as the 'rationality of irrationality.' Appearing unpredictable wasn't a flaw in your strategy. It was the strategy. Stability, paradoxically, required the credible threat of madness.
In the World
In 1983, a Soviet early-warning satellite detected what its software interpreted as five American intercontinental ballistic missiles inbound. Lieutenant Colonel Stanislav Petrov was the duty officer at Serpukhov-15, the Soviet command bunker outside Moscow. Protocol was clear: report it up the chain immediately, triggering an automatic Soviet launch before the missiles arrived. Petrov didn't. He reasoned — correctly, as it turned out — that a real American first strike would involve hundreds of missiles, not five. The satellite system was new and had a known glitch involving sunlight reflecting off clouds. He marked the alert as a malfunction and filed a report describing it as a false alarm. He was reprimanded. Not for getting it wrong, but for improperly filling out his logbook during the incident. The episode exposes the terrifying gap at the heart of MAD: the entire system depends on the people inside it behaving rationally under extreme pressure, with incomplete information, in seconds. The doctrine assumes sovereign states as rational actors, but it is always individual human beings — tired, uncertain, working with faulty equipment — who actually hold the decision. Petrov is sometimes called 'the man who saved the world.' He died in 2017, relatively unknown in Russia, having spent years trying to get his military pension restored. The doctrine he quietly defied that night is still the foundational logic of nuclear deterrence.
Why It Matters
Nuclear strategy can feel like an abstraction — a Cold War relic, a museum piece. But the logic hasn't gone anywhere. There are still roughly 12,000 warheads distributed among nine countries, and the doctrine of deterrence still governs how those weapons are held and signalled. When a nuclear-armed state makes a veiled threat during a conventional conflict, it is deliberately invoking MAD — reminding its opponents that escalation has no clean ceiling. Understanding this changes how you read those moments. The threat doesn't have to be sincere to be effective; it just has to be believed. Which means the actual danger lies not in leaders who want nuclear war, but in the possibility of miscalculation — a false alert, a misread signal, a moment where the gap between doctrine and human judgment proves fatal. Petrov's story is worth carrying around. Not as a cause for despair, but as a corrective to the idea that systems of deterrence are self-executing. They are not. They run on people.
A Question to Ponder
If a strategy only works because the other side believes you might act irrationally, at what point does performing irrationality become indistinguishable from the real thing?
Get a new one of these every morning.
Start learning with Thinkable