Not really. Postgres' planner is also statistics based. Yes that approach has problems, but the contrary, where you still use a plan that was appropriate for th

Not really. Postgres' planner is also statistics based. Yes that approach has problems, but the contrary, where you still use a plan that was appropriate for three rows with a couple million isn't good either. It's just a hard problem.