@liw@toot.liw.fi Let's make a Nash diagram: Petrov had two options: Launch (L) our Not Launch (N), and there where two options: The attack was real (R) or there was an error (E). What where the outcomes of the four possibilities. LE: He would have started a nuclear war an would be dead in 40 minutes. LR: He would have engaged in a nuclear war and would be dead in 20 minutes. NE: He would have prevented a nuclear war. And live NR: He would not have retaliated against his enemies. It was just one rocket. Retaliation would have begun when te first bomb exploted he would be dead in 50 minutes. Unless his base was the target then it would be dead in 20 minutes. As a human he dit the thing what would give him the longest time to live regardless of what the situation was. This is a post about AI in the battle field.