If you read the literature it turns out that's in some cases wishful thinking. You'd never get full yield, but no chain reaction turns out to not be guaranteed in the older designs.

Can you point me towards what you read on this outside of the article? Not doubting you it's just been like 500 years since I studied nuclear chem and I'm not sure what literature you're talking about.

A good introduction is Eric Schlosser's Command and Control. For the points here the concept of one-point safe is the idea of if a nuclear device is likely to detonate in an uncontrolled situation. A good definition I found was[0]:

> What is one-point safe?

> Apart from preventing unauthorized use, it is equally important to ensure that the weapons do not explode accidentally. For example, if it is accidentally dropped dur­ing transportation (such incidents have occurred), it should not explode. A nuclear weapon is one-point safe if, when the High Explosive inside the weapon is initiated and detonated at any single point, the probability of producing a nuclear yield exceeding 4 pounds TNT equivalent is less than 1 in one million.

A surprising number of early nuclear devices were not one-point safe.

[0] http://www.nucleardarkness.org/nuclear/nuclearandconventiona...

That only applies to most modern warheads. Older warheads had much less safety features.

W69 was designed in early 70s when they had much less safety features. W69 had a design error that made it nuclear detonation possible in the event of a fire.