However, this doesn't highlight good engineers. You know why? Because in the real world finding out the problem is many times a lot harder than coming up with a solution.

Here's a sample:

If I say to you that "I want these and these people classified / sorted, such that I want to find the people that have the best probability of being a good hire" ... I haven't actually told you the problem. I don't really know the problem myself. All I know is maybe that the hiring process is broken for some reason and I'd like to fix it, but I don't know the reasons (I can't put my finger on it).

And that's not a properly defined problem. We don't even know how to measure properly if a hire was good or not. And in the real world, that's how most problems are described to engineers - i.e. I have this pain, solve it somehow.