A tool comparison should keep the task where both answers were weak. Rankings are easier to read when the shared failure is visible.