I'm not even sure that coding _much_ faster than necessary is even required to give a 3-5x multiple on "average", let alone "worst case" developers. Some of the biggest productivity wins can be had by being able to look at requirements, knowing what's right or wrong about them, and getting everyone on the same page so the thing only needs to be made once. Being good at test and debug so problems are identified and fixed _early_ are also big wins. Lots of that is just having the experience to recognize what sort of problem you're dealing with very quickly.
Being a programming prodigy is nice, but I don't think you even really need that.
Being a programming prodigy is nice, but I don't think you even really need that.