Logout 
Portal 
Forum
Search
Members
Help 
01-14-2014, 04:28 AM
This is a tentative evaluation of EQF's forecasting graphs.
He has claimed that when a peak appears in his graphs it indicates a greater probability of a quake at the indicated longitude at some future time.
To test this idea I used his 2001-2013 graph posted on his website.
I had the lines digitized and summed the values into one line representing the total amplitude for each degree of longitude. I also eliminated the 20 degree overlap at each end.
Then I similarly summed all mag 6+ quakes in the NEIC catalog by degree of longitude and plotted both for visual comparison.
Both lines were converted to percentages of their total ranges.
I also computed the coefficient of correlation which was 0.245 indicating a small similarity, not large enough to be significant.
Clearly these graphs do not indicate an increased probability for a quake. They are random curve which sometimes match and sometimes do not match.
Roger
He has claimed that when a peak appears in his graphs it indicates a greater probability of a quake at the indicated longitude at some future time.
To test this idea I used his 2001-2013 graph posted on his website.
I had the lines digitized and summed the values into one line representing the total amplitude for each degree of longitude. I also eliminated the 20 degree overlap at each end.
Then I similarly summed all mag 6+ quakes in the NEIC catalog by degree of longitude and plotted both for visual comparison.
Both lines were converted to percentages of their total ranges.
I also computed the coefficient of correlation which was 0.245 indicating a small similarity, not large enough to be significant.
Clearly these graphs do not indicate an increased probability for a quake. They are random curve which sometimes match and sometimes do not match.
Roger