This is just my opinion based on my work in process control.
I have never yet seen even a simple system that worked the first time it was tested, and to my knowledge we have never launched a nuclear strike using the SIOP. The number of controls and checkpoints that must necessarily have been built into that launch process to prevent an accidental launch would almost certainly prevent an actual (valid) launch.
Having said that I think it’s a minor miracle that we all made it through the period 1978 to 1990 in one piece.
Again, just my two pennerth.