[mannykannot]

> but it's possible that z causes both x and y, and z's effect on x is just more delayed than its effect on y.

This is in fact the case with the barometer falling before a storm. Both the falling barometer and the subsequent rain and wind of a storm are consequences of an uneven distribution of heat and moisture in the atmosphere approaching equilibrium under the constraints of Earth's gravity and Coriolis force.