What a fun little read on a most unexpected topic. I'm certainly one to appreciate the gumption it takes to write a guide on something so subjective and ambiguous...but there's one statement in particular that I can't in good conscience abide unchallenged:
Its always best to work on the hardest aspects first. Leave polish for last. Get the main mechanics working first. So lets say you are doing an Rpg game. First begin work on your inventory mechanics, Leveling system, Save/Load feature etc.
I strongly advise against this approach. Instead, do what it takes to get a playable prototype up and running as quickly as possible: make that first area with that first enemy! I've seen too many budding designers playing around with buttons and knobs on YouTube, after spending twenty hours on a project without an ounce of gameplay to show for it. Save/Load least of all should be a priority...as a developer who has full control over his "avatar" you're likely to be manually buffing/moving him as needed throughout development anyway.