> That's extremely concerning. If we rely on MAD for survival, we should never signal weakness or unpreparedness with our nuclear arsenal. Not as much as you w

> That's extremely concerning. If we rely on MAD for survival, we should never signal weakness or unpreparedness with our nuclear arsenal. Not as much as you would think. Ballistic missiles are not the most reliable things in the world, you expect some degree of failures, especially at longer ranges. When Iran launched its massive ballistic missile barrage at Israel last year, I saw estimates that as many as half failed, either during boost phase or during the midcourse phase. And while that isn't great, it's not entirely unexpected. American ICBMs are expected to be more reliable than that, but we're probably talking about 70-80% in realistic conditions (My source is the "Iran and Israel's Missile War" episode of the Arms Control Wonk Podcast, where Dr. Lewis makes those statements. A quick google didn't give a written source). The reason this doesn't matter overly much is because the US (and other nuclear powers) have a lot of missiles. Even if your missiles have 50% reliability, if you just make twice as many missiles as you need there isn't really an issue. You still want your missiles as reliable as possible, of course, and you want your tests to be especially reliable for signaling purposes. But a failed test every now and then is not going to fundamentally alter the risk calculus.