Yes, the predictor has to set the boundaries in order to accurately evaluate the prediction. Hence the need for lats/lons. Looking at the recenteqs map, they will list out locations along with the coordinates, but sometimes the closest place is 50 miles away. Looking at the NEIC site, they use named locations such as the Banda Sea, or Alabama, or Colorado. So in the sense of being able to accurately measure whether or not the prediction was a success, specific coordinates are needed. But those specific coordinates don't necessarily have to be the exact place that the quake will occur. But then again it depends on the area your posting the prediction for. The predicted area could be within a box that could either be triangular or square. In that case the coordinates for each of the corners would be needed. Or the predictor could use a radius around an area using one specific center coordinate. Its whatever the predictor feels most comfortable with. All that really matters is whether or not the prediction is clear enough in order to accurately measure the success or failure of that prediction without any ambiguities. Dennis