I'm oddly intrigued by the phasing/extra spacey statics. I'm discovering stuff bit by bit. The bigger the object, the further dynamic objects can float or sink of course. At least, I think so. Though what I just found is that the angle of a rectangle affects it's physical shape.

Rectangles!

http://www.fantasticcontraption.com/?editId=619749 (The level I was toying with)

Rotate the square down right, and at the perfect angle, the ball falls far. Rotate it too much or too little, and the physical part rises up. It seems like there is a fixed maximum depending on the size of the static, though different angles will give different drop-depths at random.

Rotate the square so it slants down left. The ball now floats on the static. Different angles give different floaty-heights. It seems as if there are only two floaty-heights, but that just may be me being bad at judging the heights of stuff without a reference. (I could tell that there are multiple depths for the downright rectangle since the edge of the static was nearby the goal ball, making guesstimating easier. {If these two sentences in the parenthesis are confusing you even more, just ignore them, they are only explaining what I meant by "reference".})

So yea. The general consensus for rectangles is that downright=sink and downleft=float, with random heights/depths.

Circles!

They're simpler. Waaaaay simpler. Bigger circle=more drop-depth. The only weird catch circles have is that at multiple-of-45 angles, things won't drop into them. About 22.5 degrees is the optimal drop-depth. As far as I can tell, you cannot float on circles.