Hi Shan. Look at each quake that is listed there. I use a 5-day time window for a single forecast with a 35Km radius. If you apply that time frame to each forecast you will see that it comes out nearly the same. The last quake forecasted in the sequence is on 05/09/2002. The window for this quake would be from 05/07 2002 to 05/11/2002. That makes it 24 days for the time of occurrence from the day I made the post. I added 3 extra days because of the time frame from the time I posted to the time period of this quake is a little long there can be a change in the data during that time period which could influence the data. Take Care…Don in creepy town