EQF,

Cute web page. The answer to my question above from this page is no, you don't include the twice-daily tides, the biggest of the tides. All the charted lines have biweekly periods except the magnetic measurements.

One of your basic conclusions, M8+ earthquakes tend to happen (10 out of 17) when the Sun-Moon angle is small, (is this right, Roger?) would rule out tidal stress or strain as a factor. The tidal stresses are nearly identical whether the Sun or Moon is in one particular longitude or 180 degrees away.

The pattern is unexpected, the criteria is ad hoc ("in the range from -40 to 35 degrees in alignment"), the pattern was defined after looking at the plot, and the pattern fails to appear for earthquakes less than M8. The small chance of seeing 10 out of 17 at random (1 in 2000) is small but not small enough to convince a skeptic like me. If you had chosen the most likely pattern (for example, when stress is high) and a simple measure (dividing the data into even halves) such a significance would be meaningful, but the number of degrees of freedom in your analysis are enormous.

You expect "very clear, impressive, and important results" from this method. Yesterday alone I learned of several ideas requiring only well-documented physics, based on statistically-valid studies, although I'd judge neither are likely to to work, let alone work well enough to apply to seismic hazard reduction.

Given that you have neither a physical model nor statistics to show there are any real patterns in your data, your statement of my guilt should people die in an earthquake is ludicrous and offensive.

John