An eclipse calendar as a testable cycle scaffold

Lock the calendar before the signal

Editorial note: TradersWeek reads a publicly dated alignment calendar as a sampling grid, not as an omen. Event times are locked first. Price is then inspected for inversion at the event, the half-span, or the third-span. Only after that inspection may later filters accept the next mark as a signal.

In the archive workflow, eclipses are treated as Sun-Moon-Earth alignments that occur when the Moon is near the ecliptic. The lunar orbit is inclined about five degrees and crosses that plane at two nodes.

Two mean intervals and a physical sketch

Two mean lunar periods are specified. The draconic-interval is the mean time from one lunar node crossing of the ecliptic to the next corresponding crossing, given as 27.212221 days. The synodic-interval is the mean time from one Sun-Moon alignment, new or full, to the next, given as 29.530589 days. Ordinary lunations occur off the ecliptic.

The mechanistic sketch treats orbiting bodies as raising tidal swirls in solar gases that modulate radiation and solar-wind particles. Those particles then shape Earth's magnetopause, magnetotail, and polar ionization.

Solar eclipses are described as the Moon sitting outside the dayside bow wave and blocking inbound energy at alignment. Lunar eclipses are described as the Moon sitting in the nightside plasma sheet and briefly stopping reflected ionized particles. A near-eclipse is an approximate Sun-Moon-Earth alignment that falls short of a full eclipse but is treated as a related energy-flow event.

How past eclipses were classified

A review of eclipses from 1900 onward on a major industrial average classified each event as no effect, a move under 10 percent, or a larger move. The spreadsheet stored move size, duration, time of year, and other alignments.

That classification labeled 51 percent of all eclipses as coincident with a move, split as 18 percent larger and 33 percent smaller. Clustered events exceeded a 90 percent labeled-effect rate. An eclipse-cluster is a compact group of solar, lunar, and near alignments handled as one seasonal window rather than isolated dates.

Solar events were tabulated more than 10 percentage points more often in the combined-move class than lunar events. Near-alignments were treated as potentially as relevant as full eclipses.

Cycle lines and inversion-points

The working procedure marks eclipse times on a price chart, connects them with a straight cycle line, inserts inversions at midpoints or third-points, and uses those marks with geometric trendlines to anticipate the next event.

A dominant-cycle is a repeating interval inferred from dated event marks on an ordered price series and used to project the next inversion. An inversion-point is a hypothesized cycle turn inserted at the midpoint or third-point between two marked events. A geometric-overlay is the set of straight cycle lines and trendlines drawn on the same chart scale as the event marks to test a repeatable geometric turn hypothesis.

Two illustrated sessions

One illustrated session on 7 August 1990 showed an opening decline of more than 100 points that halted near 10:15, with a lunar-eclipse peak listed at 10:10 that morning.

A July 1991 Nikkei example was presented as following a half-eclipse-rhythm, turning at near-eclipses and at the halfway marks between them. A half-eclipse-rhythm is a working hypothesis that turns occur at eclipses or near-eclipses and at the midpoint between them.

Later gates, not first causes

Editorial note: the archive stored time of year and other alignments with each classified event. TradersWeek treats that record as a seasonal-filter, meaning time of year used with other concurrent alignments as a rule input that can confirm, weaken, or veto an event-timed signal.

Editorial note: the geometric-overlay is the second gate. Straight cycle lines and trendlines stay on the same chart scale as the event marks. They do not replace the dated calendar. They only accept or reject the next hypothesized turn after the inversion test has been stated.