The article Gowtham linked to seems to be the one you want. Basically (if I understood correctly), if the material is already hydrophobic (water is more attracted to itself than the material), the surface tension of the water prevents it from filling small empty spaces in the surface which will remain filled with air. Thus the contact surface and attractive force between water droplet and the surface will be very small.
The pattern itself is quite simple. It is simply very small empty spaces the walls of which are also full of even smaller empty spaces. This together with the material means that barring some outside pressure forcing the water into the spaces, surface tension will keep it out. Gravity will not be enough.
And in fact if you force water into the surface it will try to bounce back as long as water is less attracted to the original matter than it is to water. Water trapped in the small spaces will be more attracted to the rest of the drop than it is to the surface and if the drop is small enough and the space is small enough, this attraction wins over the gravity trying to force water to the surface.
Hope this is in the correct direction AND at least somewhat understandable.