[breatheoften]

The premise in the thread seems to be that after the first nuclear skirmishes, the only optimal strategy is to delay all-out response on the hope that initial firings were a fluke, an error, or a mistake. And that the ability for a delayed response is assured, so there's no point in rushing to press the annilihation button.

It may be true that delay would be the best option for maximizing probability of survivability in that scenario, however I do not think the capability exists to actually do this -- it might not even be possible to implement this strategy. I don't think it's possible to construct a control system that would manage the diverse array of retaliatory launch systems in a way that both preserves "guaranteed total annihilation of aggressor" and "arbitrary delay for retaliation after first strike"

[3pt14159]

You only need a very small number of doomsday devices in order to allow arbitrary delay for retaliation after first strike. Ten sleeper agents with extreme bio weapons in the likely aggressor country, for example.

There is no purpose in rushing to doomsday and a great deal of interested parties in not doing so.