Well, let's consider it this way.

An earthquake is a brittle fracturing event, or to put it simply, rocks breaking.

Consider some other sort of brittle fracturing event. Imagine you take an ordinary china saucer and grip one side of it in a vice. Now level a piston of some sort against the other side of the saucer so that it is pressing down on it. Gradually increase the pressure of the piston. Eventually you reach the breaking strain of the saucer and it snaps.

Now: how do you predict when it will snap? If you knew in advance the saucer's breaking strain and the rate of force increase in the piston, you could *calculate* the time of breaking, but trying to *predict* it not knowing these two variables is another matter. What sort of precursors are you going to get?

An earthquake is similar, just much larger and more complex. So the answer to the question frequently levelled at seismologist, "why can't you predict earthquakes" can be "why should earthquakes be predictable?". Other types of breaking event are not.