What sites would they be targeting? Generic "just give me anything"? Whenever I check regular sites on IA, the coverage is spotty -- they'll have the homepage and a few important pages, but it quickly fizzles out.

Very understandable, you can't store all 15000 pages of any random website and update them etc etc, but that makes them pretty useless for indirect scraping because you usually don't want a tiny taste, you want everything.