Yep. That MAD prevents nuclear strikes is an absolute article of faith at this point, used to reject unease outright; not an objective, practical calculus. Like all natural phenomena there will be exceptions and deviations. A nuclear incident is inevitable.
The real question to my mind is whether the benefit of nuclear deterrence suppressing non-nuclear conflict is greater than the cost that will be incurred from MAD failures/exceptions. But that's an almost impossible question to answer today. Still, it's a more honest perspective. I suspect we don't look at in this way because the public would almost certainly decide (perhaps irrationally) that they're far more afraid of a nuclear strike than the slow slaughter of conventional warfare, creating pressure to discard nuclear weapons entirely.
The real question to my mind is whether the benefit of nuclear deterrence suppressing non-nuclear conflict is greater than the cost that will be incurred from MAD failures/exceptions. But that's an almost impossible question to answer today. Still, it's a more honest perspective. I suspect we don't look at in this way because the public would almost certainly decide (perhaps irrationally) that they're far more afraid of a nuclear strike than the slow slaughter of conventional warfare, creating pressure to discard nuclear weapons entirely.