The formal New-Moon test opens - and how it differs from the informal idea

Reconstructed archive entry. It describes 11 September 2026, and was published on this site on 22 September 2026. It was written up from contemporaneous records after the fact — this site did not publish it on the historical date.

The formal prospective test begins at the New Moon, 11 September 2026, 03:27 UTC
(computed exactly: 03:26:55).

Two things must be kept apart, and the separation is the point:

1. The paper's protocol (formal, pre-registered).
Does a strict local low form in T0:T+3 - 11 to 14 September inclusive? The
definition was fixed in advance and is defined on the daily LOW, not on
closing prices: LOW[t] < LOW[t-1], LOW[t-2], LOW[t-3] AND
LOW[t] < LOW[t+1], LOW[t+2], LOW[t+3]. Strictly less-than on both sides.
This is the claim being tested.

2. The informal private NM+0.4-day timing idea.
A separate, private, much more precise timing notion. It is not the paper's
protocol, it was never pre-registered, and it must not be allowed to stand in for
the formal test if the formal test fails. Conflating them after the fact would be
exactly the kind of goalpost-moving the research constitution forbids.

The New-Moon geometry strength recorded in advance was 79/100. The formula
implemented in this codebase scores the same event 69.2/100. Both are kept.
Neither is evidence that a low must occur - strength is context, not a signal, and
it does not enter the Pattern Fit Score.

An observation, not a prediction. Recorded as part of an ongoing public experiment and not investment advice.