Those interested in foreshocks as a predictive measure might like to look at a paper by Paul Reasenberg which I came across recently: Foreshock occurrence rates before large earthquakes worldwide, in Pure and Applied Geophysics, vol 155, pp 355-379 (1999). One of his results is that that the general probability that any m >= 5 event will be followed by another larger event in the same place within seven days is 7.5%.