Moon phase calculator methodology

This moon phase calculator methodology page states exactly how every phase on the site is computed, how two phases get assigned to a compatibility category, and which parts of the result are measurable astronomy and which are astrological interpretation. Every figure here can be checked against an independent ephemeris.

The astronomy

A birth date is read at noon UTC and converted to a Julian Day. Lunar longitude is computed from the truncated ELP-2000/82 series printed as Table 47.A in Jean Meeus, Astronomical Algorithms (2nd edition, 1998), chapter 47, using the 59 largest periodic terms plus the planetary and flattening corrections. Solar longitude comes from the low-precision series in chapter 25 of the same book.

Elongation is the Moon's longitude minus the Sun's, normalised to 0 to 360 degrees. The illuminated fraction follows as (1 - cos elongation) / 2, and the named phase is the 45-degree bin the elongation falls in, with each phase centred on its exact moment: New Moon spans 337.5 to 22.5 degrees rather than starting at zero.

Lunar longitude from this series is accurate to roughly 0.02 degrees, which places the illuminated fraction within a few tenths of a percent of a full ephemeris. That margin is far smaller than a 45-degree phase bin, so the named phase is wrong only for dates within a few hours of a bin boundary, where a birth time would be needed to settle it anyway.

Why there is no new-moon epoch in this calculation

Many moon phase tools count 29.53 days forward from one reference new moon. Individual lunations actually run from about 29.27 to 29.83 days, so the errors accumulate: a fixed cycle can be a full day out within a few decades and name the phase before or after the right one. Computing the Moon's position for the date itself avoids that drift entirely, and it is the main reason two calculators disagree about the same birthday.

The mean synodic month of 29.530588861 days appears in only one place here: converting an elongation into a human-readable "days into the lunation" figure for display. It is never used to determine the phase itself, so no epoch date is needed and none is stored.

Known limits

• Dates are read at noon UTC. A birth in the hours around an exact new, quarter or full moon can fall in the neighbouring bin when read in a different time zone.
• The illumination figure is geocentric and ignores the small topocentric correction for the observer's position on Earth, which affects the percentage by well under 1%.
• The input range is 1900 to 2099; the series is accurate outside it but untested here.
• No birth time is accepted, by design. Accepting one would imply a precision the phase boundaries do not support for the great majority of dates.

How a pair is assigned to a category

The eight phases are indexed 0 to 7 around the wheel. For any two phases the separation is the smaller of the two distances between their indices, giving a value from 0 to 4, which is multiplied by 45 to get the angular separation. That number alone determines the category, there is no scoring, no weighting and no percentage, because the underlying framework does not support one.

Two of these names are exact: 0 degrees is a true conjunction and 180 a true opposition. The other two are approximations, since an eight-point wheel produces 45 and 135 degree separations rather than the 60 and 120 degrees of a real sextile or trine. They are labelled sextile and trine here because that is the vocabulary readers arrive with, and the exact degree figure is shown alongside every result so nothing is hidden by the label.

Some compatibility systems list a sixth category, aversion, for phases forming no aspect at all. An eight-point wheel produces only five distinct separations, so a sixth category would have to be invented rather than derived, and this site does not use one.

The astrology layer

Phase personalities and pairing readings follow the eight lunation types in Dane Rudhyar, The Lunation Cycle (1967). Pair categories are named for the chart aspect each wheel separation sits closest to.

No scientific evidence links the moon phase on a person's birth date to their personality or their relationships. The astronomy on this site is checkable; the interpretation built on it is an astrological framework and is presented as one.

Editorial policy

• No invented social proof. There are no review counts, star ratings, testimonials or "people calculated this today" counters anywhere on the site, because none would be real.
• Astronomy and astrology are labelled separately on every page that contains both, rather than blended into a single confident voice.
• Errors in the astronomy are bugs. If a date returns a phase that disagrees with a published ephemeris, that is a defect and gets fixed. Report it through the contact page.
• Nothing is stored. Birth dates are processed in the browser and never transmitted, as set out in the privacy policy. The site terms cover what the readings are and are not.

Sources

• Meeus, J. (1998). Astronomical Algorithms, 2nd edition. Willmann-Bell. Chapters 25 and 47 supply the solar and lunar longitude series used here.
• Rudhyar, D. (1967). The Lunation Cycle. The source of the eight lunation types behind every personality description on the site.
• Rotton, J. & Kelly, I. W. (1985). "Much ado about the full moon: a meta-analysis of lunar-lunacy research." Psychological Bulletin 97(2), 286 to 306. This meta-analysis pooled 37 studies of moon phase against psychiatric admissions, crisis calls, homicides and other offences. Lunar phase accounted for no more than 1% of the variance in the behaviour examined.
• Foster, R. G. & Roenneberg, T. (2008). "Human responses to the geophysical daily, annual and lunar cycles." Current Biology 18(17), R784 to R794. doi:10.1016/j.cub.2008.07.003. A review of which geophysical cycles measurably affect human biology and which do not.

The 2 research references above are cited for one claim: no reliable lunar effect on human behaviour has been established. Neither is cited in support of anything the astrology layer of this site says, because no study supports it. Both were checked against the published record in September 2026.

Checking the output

Any result here can be compared against an independent ephemeris. As reference points, the calculator returns 0.1% illumination for 6 January 2000 and 99.9% for 25 January 2024, both of which carried an exact new and full moon respectively.

Last reviewed 9 September 2026. For the site's purpose rather than its arithmetic, see about this site, or return to the calculator itself.