The problem is that you trade the virtual certainty of periodic destructive wars for a small but non-zero chance of an utterly apocalyptic war. The big question I don't see people asking is, what probability makes this tradeoff worthwhile? One in a hundred per year? One in a thousand? One in a million? Based on what I know about how these systems work, and the history of near misses, I suspect the probability is closer to one in a hundred than one in a million.