The title of this post says "The Golden Rules" but what I learnt so far is that there are no golden rules to making a successful title. There are no set parameters that you can get right, they vary greatly depending upon your game idea, the genre, the platforms you're targeting, the people you're targeting and so on. But here is what I did learn, which applies to a variety of games:
Fast Iteration and Pure Design Time
Giving independent design time, eliminates the intervention of game engineers. Questions related to how a feature will be implemented and what would be its in-depth details is put off for later, leaving time to focus on deciding the gameplay elements and experience design.
Using tools such as Flash, Game Maker etc. that would require minimal code, and probably will be usable by non-programming member of the team can be of help in the prototyping phase. This prototype should be discarded when the programmed mechanic or feature has been finalized. Reusing hasty and inefficient code that was meant as a prototype can cause problems later on in the project.
Once the prototyping phase is done, start closing the game design to changes; making too many changes during the later stages of a project wastes time. Furthermore, don't reinvent the wheel. If you can get code, plug-ins and assets for a store, assess the cost and get them, quite often you end up saving a lot of time.