For fun
Chinese gender predictor
The lunar chart your family probably has a copy of. It is folklore, it has been tested against 2.8 million births, and it performs exactly as well as a coin — all of which is on this page rather than buried under it. What we do take seriously is the lunar arithmetic, which almost every free version of this gets wrong.
Two dates
The chart needs the mother's lunar age and the lunar month of conception. Both are computed here from the real lunar calendar rather than by adding one to a Western age, which is the part every free version gets wrong.
Disclaimer: entertainment only. This is not a medical test, it does not detect anything, and it should not inform any decision of any kind. An ultrasound or NIPT will tell you accurately and your clinician will arrange one. This page reads in one direction only — it answers what the chart says about a conception that has already happened, and there is deliberately no way to run it backwards to pick a month.
The part that is actually hard
The chart needs a lunar age (虛歲) and a lunar month. Lunar age is not your age plus one: you are one at birth, and you gain a year at each Lunar New Year rather than on your birthday. So it runs one or two ahead of a Western age depending on the time of year.
Somebody born in mid-January 2001 was born before that year's new year, so her lunar birth year is 2000. In June 2026 her lunar age is 27 while her Western age is 25. A calculator that adds one to the Western age puts her on row 26 — a different row of the chart, and a different answer for half the months of the year. Both dates here go through the real lunar calendar.
Which version of the chart
Published copies of this chart differ from one another in a handful of cells, which is what happens to a table copied by hand for a couple of centuries. Ours is transcribed from a single published source, and when it gives you an answer it also prints the row it used with your month highlighted.
So if your family has a copy that disagrees, you can see exactly which cell differs instead of being told one of you is wrong. Given the chart's measured accuracy, neither of you is.
It only runs one way
This page answers what the chart says about a conception that has already happened. It has no control for finding the month that yields a particular result, and the engine behind it exposes no function that would answer that — there is a test asserting that none is ever added.
A chart that predicts at chance cannot select for anything anyway. But building the tool in the direction that only reads seemed better than relying on that.
