I finished Plutarch's Fall of the Roman Republic and he'd describe battles where Caesar or Pompey would kill 30,000 enemy troops and lose fifty or so of their own. How accurate are these figures? I'd imagine that during a battle, one line would break into retreat while the other would hold up, neither one taking very many casualties during the battle itself but most of them coming during a retreat. Am I on the right track?
My recent response to the question "In ancient battle, what was the casualty ratio of combat to rout?" might be of interest. As you can read there, you are on the right track with your idea.
Unfortunately I'm not familiar enough with Plutarch or the battles of the first century to give a detailed answer on the specific numbers you mention (figures in ancient sources should always be treated with caution).