EQF;

Suppose you predict a quake of mag 7+ within a week-long window within a 500 km radius of San Francisco.

The Jones method divides the time span of the database into week long consecutive windows of time. In 30 years there are 30*52 = 1560 windows.

You sort out all the 7+ quakes within 500 km of SF in the database and drop them into the correct window. Suppose ther were 20 such quakes but 3 of them were in the same window as other such quakes. That means there were 17 windows with 1 or more 7+ quakes out of 1560 windows.

So the odds are 17/1560 that a window chosen at random will contain a 7+ quake within your circle.

In this example the Jones probability is very close to simple probability based on percent of time vs percent of quakes but if you were interested in smaller events the presence of aftershocks would bias simple probability incorrectly. The Jones method eliminates most of that bias.

Roger