Roger: Thanks for your response.

Earthquakes listed on the Central Weather Bureau here in Taiwan often don't make it to the USGS site. No doubt it is the same worldwide. Taiwan and Japan have world class seismic monitoring. It seems it shouldn't be all that difficult for the USGS to get the data quickly. As for other parts of the world, who knows. At one point, I emailed the USGS about this and they replied telling that they only claim accuracy of events larger than M5. Over the last weekend, the USGS "Near Real Time" event list was not updated for fifty hours straight. Isn't it likely that they missed a lot of events during that period?

There still must be some way to use calculations to locate the most probable populated area for a large event. Have you ever tried to list all events over some magnitude and see if there is any trend. Doesn't the USGS use this type of information to base their 30 year windows on? Perhaps with the combination of historical seismicity and trends in locations the 30 year window could be reduced??

Even if the event found to be the most probable was not the next big one, odds are it would be one of the next big ones. Do you think the USGS actually can narrow it down to less than 30 years but leaves a big safety margin of error, or is 30 years the best that can be done considering time on a geologic time scale?