Earlier you say that predictions should follow the scientific method. But the scientific method demands lats and lons. Now you are saying to use city. You aren't following the scientific method anymore?

When a person uses a city, that increases the size of the coverage area when not using a specific lat/lon. In the sense of determining whether or not a quake is a hit, where in that city would you start measuring from?

Jim Berkland posts predictions for within 140 miles of a city. When he determines whether or not the prediction is a hit, he bases it on that cities boundaries. Not by that cities coordinates. Some cities, like Los Angeles, are 20 miles or more from one boundary to another. It makes it very hard then to determine whether or not an event is a hit because most maps don't have the lats/lons of the boundaries for each city.

Another aspect of this is that it wouldn't be site specific when a person is not using specific lat/lons. For instance, if you predicted a quake in Los Angeles, are you talking about down by San Pedro or are you talking about around Northridge. A quake in one of those locations wouldn't affect the other location very much.

Antonio should continue using lat/lons as he has been. He could add in that these coordinates are close to city x. But then again, the closest city may not be on any map and one that is on a map could be 10's of km away from the specific location.

Dennis