This is why it’s good to intentionally go between the micro and macro and define some design goals or pillars. The latter give you a razor by which to judge the game and the former stops you getting stuck in the weeds.

I broadly agree with the Romero quote, getting the game playable as quickly as possible and keeping it playable is easily the most effective way of crafting a game because it enables you to routinely playtest and understand your progress. A key element of that is making the game legible and for that you do need to spend some time on “polish” because it’s an intractable element of the whole.