>While I appreciate how hard it would be to implement properly

Isn't that part of what it means to acquire a patent? Trying to get a patent on something that I don't yet know how to actually implement should be denied, right?

Nope. It's actually best practice to get one even without an implementation because usually the threat of a lawsuit is enough of a chilling effect to keep the competition guessing.

Of course the risk is that if someone else challenges you on it, and you can't provide the implementation, you can have your patent invalidated.

At least that's my understanding of the tactics of patent warfare.