Hi Michael. Comparing a true prediction against a random prediction would show how well a true prediction is. If the list of true predictions has a greater number of hits it could be said the person making the predictions can indeed make an accurate prediction. If on the other hand it is equal, or worse then the random list it would show that the true prediction, or the person making the predictions did no better, or worse then random choice.

I do this for the area I live in. However I also include earthquakes that were felt in this area for which a prediction was not made. I have 7 correct forecasts for the SJB/Hollister area, but there have been many more quakes for which a forecast was not made. I use the data site at SJB that covers only a small portion of the San Andreas and Calaveras faults. We have 17 active faults in the area so it’s very easy to have a quake for which there will be no data seen. The same applies to the Parkfield area. I have made one correct forecast for that area. About 2 years ago I made a forecast for a quake near Middle Mountain. The most recent 3.4Ml occurred in the same area, but this time there was no indication prior to the quake.

I have made 11 forecasts for the Bay Area over a two year period. Not one of them was good. As a matter of fact the random list I have did better then I did. Even with a perfect record of misses I’m learning. Take Care…Don in creepy town.