Which is why the nuclear test ban is, IMHO, a bad idea. A nuclear deterrent must be credible to be effective. If an adversary comes to believe that, say, 80% of our warheads are duds and most of theirs work, the logic of retaliation may come to favor a first strike.
The logic behind test bans (in conjunction with numerical caps on warheads) is to create uncertainty in the reliability of one's own arsenal to discourage either side from contemplating a first-strike.
I'm not sure about that. What if you're convinced that your brilliant scientists have created working warheads while you think the enemy's dolts haven't been able to keep their arsenal working? What if your enemy thinks the same thing in reverse? I think there's always a temptation to overestimate one's own capability and underestimate the sophistication of others.
