Once development begins on a feature, game developers often get drawn into a sunk cost fallacy that compels them to keep working on that feature even when it's failing. Cutting those features is hard, but often the right choice to make the game better.