The issue is that every time you make a video game, it's made just slightly differently enough (different mechanics! different art style! changes in tech while you made the last game!), that quite a lot of the stuff you "know" just is not relevant any more.

Yup, making several games just makes you put your hands up and go "I dunno, let's try something..." whenever a unique new problem pops up first thing in on a fresh project.

The key is understanding the underlying 'broad stroke' principles and hoping for the best with the little things like making the little man run around.