The difference between "good" and "great" lawyering is practically invisible before the fact. You only notice its effects after, and by then it looks simple and obvious—but it never was, which is why few can do it that well.
I suspect this observation has applications in many other fields too.
I would add that, most of the time, good lawyering is sufficient. Client-significant outcome differences between good & great lawyering occur in a minority of instances. That clause of your contract that might have been more exquisitely drafted probably won’t hurt you.
This is why there’s so much mythos around the rockstar programmer. A lot of the work is the regular grind, code maintenance, updating libraries, monitoring, implementing new algorithms. But when a well designed system comes together it’s magic.