Petra and Don,
I thought you might like to see a few statistics on Don's prediction.
Because the prediction was rather vague, it does not score high, however.
The prediction stated:

"Now lets see if there is a quake in the range of 3.5Md to 5.0Md north of the Bay Area.
Some where around Eureka, Lake Pillsbury. Give it about 3 to 7 days. "

According to NGDC earthquake records for California, there have been 1523 earthquakes
with Md 3.5-5.0 in the past 35 years in the region north of the Bay (38-42N 120-126W). This averages to a daily chance of such an earthquake of about 1 in 10. Don said the earthquake
should occur in about 3 to 7 days (a window of 5 days). The probability of a correct
prediction by chance is about 5 times the daily probability, or about 1 chance
in 2. The earthquake actually did not occur in Don's window, but outside. If we expand
the window to 1-7 days (7 days window) the odds of being right by chance are even
greater.

What could Don have done to improve his forecast? The best approach would have
been to make several forecasts - say within 50 km of Eureka and within 50 km of
Lake Pillsbury, rather than take in the entire region of Northern California. Giving a
shorter time window would also help, but may not be possible using the indicators
Don used.

There is no reason to believe that Don's techniques do not yield useful results given
this one example. By refining his prediction and technique, it is possible that
useful results could be obtained. The fact that Don posted this prediction within
a day of a Ml 4.1 in the stated area certainly suggests he may be on to something.
On the other hand, it is important not to be fooled by optical illusions or in this
case statistical illusions. Care should always be taken in interpreting any prediction,
being careful not to jump to hasty conclusions without first checking the data.

The o
in this m