You can see from these latest criticisms that my mention of Copernicus was more about learning from his error in assuming circular motion of planets than any claim to genius on my part! The issue remains that an average random 345 days in the study period has 1142mm of rain, but when these dates are sorted into moon-planet cycles the groups show surprisingly large spikes, as much as 600mm above or below average. When the data is aggregated in this way this variance is large. And, the point which the helpful comments on my method did not address, two big data spikes occur on adjacent days of the moon-sun cycle, at the first quarter. On the face of it this result looks much bigger than could arise from chance. I have studied some statistics, but not enough to know how to determine the probability that these results are from chance alone. The binomial method is obviously wrong in assessing outliers, but I would have thought it gave a good indicator of probability of results closer to the mean, certainly enough to show that these results are highly significant.

And regarding the small effect of Uranus, my point was that the permanent rhythm of this weak force is likely to be amplified and entrained over the extremely long period in which it has been present. There are no stadiums full of people which have been orbiting the sun with mechanical constancy for nearly five billion years.