[contravariant]

Maybe but 1) that's hard to prove and 2) it's hard to achieve unless you're looking at some sub-problem where the quantum and classical algorithm actually do scale differently.