If only earthquake modelling was so simple as to say you were overdue for an earthquake simply based on accumulated stress. Unfortunately, the Gutenberg-Richter law is not prescriptive, but a time independent statistical observation which assumes independent events, which is why saying you're due for another one is very much like gambler's fallacy. Of course we know that there are underlying physical processes which make means there should be some underlying dependence to be found, and there are attempts at time-dependent models out there, but for our purposes independence works.

If you're interested in the statistics around recurrence times, can I recommend this [1]? It's fairly well written and clear and it deals with this issue fairly comprehensively and directly.

[1] - http://www.crm.cat/en/About/People/Researchers/ACorral/Docum...