Despite a total of 5 misses, the experimental astrological forecast model is holding up very well statistically versus the random dates control group forecast model.
The astrological forecast model observes 6 EQ’s occurring out of the 8 that have been predicted so far, for a simple score of 75% correct.
6 ‘hits’ out of 8 ‘trials’ given a frequency of .359 eq’s per day yields a binomial probability of (p) .027, or just 3 chances out of 100 of getting a 75% score on this forecast.
When the ‘no-quake’ days are factored in, the astro-model gets 15 hits out of 20 trials (again, a simple score of 75% correct) for an impressive binomial distribution value of (p) .005, or just 5 chances out of 1,000 of getting a score of 75% correct.
Lastly, a simple EQ days v. no EQ days analysis reveals a pretty fair correlation value of (r) .47 .
A correlation ratio of .47 is not something one would normally write home about, but given the presumed total unpredictability of earthquakes, these results are worth noting.
On the other hand, the random dates forecast model is performing just as one might expect: according to random chance.
The control group scores 4 EQ’s days out of 9 EQ days predicted, for a simple score of 44%, which is pretty close to the expected daily frequency of M6+ earthquakes : (f) .359 or 36%
The binomial distribution for 4 ‘hits’ out of 9 ‘trials’ yields a probability for the control group forecast of just (p) .23 or in other words the control gives nearly 1 in 4 chance of a correct score of 44% .
Finally, the correlation ration for the random dates model is a mere (r) -.01 , which is virtually no correlation at all, as indicated by the widely separated divergent trend lines.
These results, while still just a small sample and a short time frame, are very powerful evidence of astrological influence over one of the most damnable natural disasters plaguing mankind. I shall continue to ‘tweak’ the model, and if it continues to succeed, I will work to bring this matter to the attention of ‘the powers that be’.