My usual approach of getting something really, reeally right is the following: First, I exagarate doing way too much or way to little of it.Then I adjust from both sides, to get it right. Not like a binary search where I would try the center value, but rather moving those to extrema closer and closer to the center. E.g. trying to get the right shade for a UI item, I'd start with black and white. If I started with grey, I might be ok with my solution too early, never noticing that a little darker would have been better. Ok, in this case the whole process has to do with the fact that I am a zero as far is art is concerned, but the approach does work for me in general. After trying to design games from a very gamedesign-theoretic point of view, putting all the focus on mechanisms (rock, paper, sciccors, etc.) I lately found that there is a totally different approach that seems promising: Gamedesigners will tell you that your vision is all about the feeling of the game and that you have to keep that vision up throughout the process. So instead of building a design concept around the vision of the feeling I wanted to evoke in the player, I thought I could as well 'collect gameplay situations' that evoke this feeling. Aiming for a lovely lost-in-space-and-time feeling I started to prototype some game around some cute iconic rocket that is attracted to planets by their gravity and that can land on them and take off again. My working title for this game to come is 'Space Odyssey' and I am curious where putting together pieces of nice stuff will lead me..