You have probably noticed that a lot of earthquake predictors state their percentage of correct predictions. I understand Jim Berkland claims 80%. As I've often stated, one cannot evaluate a predictor in this way. What if his/her prediction is for an event, such as a very small earthquake, that is almost guarenteed to happen?

To illustrate this, I am making a "prediction" for each 7-day period starting today for California and part of Nevada. The latitude and longitude limits are: 32.5N to 42.1N and 112.5W to 126W. I am "predicting" that there will be at least one event of magnitude 3.3 from 2000/10/20 0:00:00 UTC until 2000/10/27 0:00:00 UTC. By looking at similar windows since 1960, I believe I will have at least one hit about 92% of the time.

Let's let nature do its thing and see what happens. I will try to report the results here each Friday.