Chris,

It's fun to look at correlations. The most fun, however, is when people can check their own correlations, clearly explain what they see, make a prediction, then publicly test it. I view my own research into earthquake-tide correlations that way, except I don't need to forecast, since the tides and the earthquakes are a matter of public record.

Several of us are investigating geoForecaster in a superficial way, and finding troubling problems, such as redefinitions of scoring that are not mentioned, errors in calculating hits, the proprietary nature of the algorithm, and especially a scoring system in which most probable earthquakes score highly.

I guess my point is that the only way geoF can be fairly tested at this point is by catching all the predictions as they are made, and seeing if they are useful. Some are starting on this route. Unless they posted thousands of predictions, or the method is much better than it seems so far, it would not benefit from a test of this boards archives.

If the merits are not cleared up after a couple of months, the assessment may wind up with CEPEC or NEPEC, in which case it would be formal, protracted, and definitive.

John