Why your moon sign changes between apps

Two surveyors measure the same field a week apart. Every distance matches, every bearing too. Then the office lays one map over the other and the well is in two places. Neither man erred. One measured from a granite block at the roadside, the other from an iron pin forty metres west. They counted from different stones.

Astrology software has this problem. Picture a ring of 360 degrees drawn around the sky, along the path the sun appears to walk through the year. Every app agrees where the moon sits on that ring; they mostly read the same NASA data. The argument is about where zero is.

There are two ways to fix zero. One pins it to the seasons: the March moment when the sun, moving north, crosses the line straight above the earth's equator. The other pins it to the stars, and every Indian tradition uses the star version.

The two zeros are not in the same place, because the earth's axis wobbles like a slowing top. The seasonal zero slides backwards around the ring by about 50 arcseconds a year, a degree every seventy-two years. That accumulated gap is the ayanamsha, and converting between the two is one subtraction.

How big is the gap? That depends on when you think the two zeros last matched. Lahiri says 285 CE, and Indian official calendars have followed that since a committee reported in 1955. Krishnamurti says 291 CE. Six years apart works out to about five arcminutes today.

Five arcminutes is tiny. The essay builds a chosen example anyway: a moon landing three arcminutes short of a sign boundary. Lahiri prints Leo. Krishnamurti prints Virgo. One moon, one instant, two headlines.

Two things worth noticing. The nakshatra does not change; all that moves is which of its four quarters the moon sits in. And five arcminutes of moon is about nine minutes of clock time, so the choice does the same damage as a birth time written nine minutes off.

The essay refuses to say which is right. Arithmetic cannot settle it, because arithmetic is not in dispute.