First, I want to thank Petra for her use of the EPR even though it is not really
designed for ear tone type sensitivity. It is important to use the tools available
so that consistent evaluation of predictions can be made.
Petra's prediction WAS correct in two aspects - an earthquake of Ml 4.5-6.3
did occur within the 10.5 day window she specified within the northwest region.
But as with all other predictions, it is important to examine the statistics
to see if this is really statistically unusual forecast - in other words, what are
the chances Petra would have been successful even if the ear tone she heard
did not indicate a coming earthquake.
To do this it is necessary to look at how often earthquakes occur in the
region in question. Since Petra acknowledges that the Queen Charlotte
Islands do not fall within 75 km of Bremerton, yet believes that the ear tone
she heard was a genuine precursor to the QCI earthquake, the area from which
such tones can originate must encompass nearly the entire region of Oregon,
Washington, off their coasts, Vancouver, B.C. and QCI. So I have looked at
the number of events this year (2001) in the region 55N - 40N and 145W to 120W
within the magnitude range 4.5-6.3 stated in Petra's prediction.
During 2001 there have been 36 events which fall in this area and magnitude
range - on average about 1 every 8 days. Of course many of these were in
the cluster off the coast of Vancouver last month, so it is better to use
the Jones algorithm. Using that I find that 20 10.5 day periods have had at least
1 such event during 2001. There have been 27 such 10.5 day periods in 2001,
and 20 of them saw a Ml 4.5-6.3 earthquake in this region. That's pretty good
odds of being successful.
I don't want to discourage anyone with this kind of statistics. What I am trying
to show is that if a large, seismically active area is used, the chances of
a successful prediction are pretty good no matter what method you use.
To really show anything statistically, a much smaller area is needed than
"the northwest". Of course, ear tones do not come with a subliminal message
whispering "this earthquake will occur at 48.233N 122.465W with magnitude
3.46", and the statistical method I have used here may not apply to the type
of information sensitives get about earthquakes, but I don't know any other
way to evaluate predictions besides statistics.
If anyone out there has a more suitable method, I would sure like to hear
about it.