Is the Chinese gender predictor accurate?
No. Researchers checked the chart against 2,840,755 Swedish births and found it agreed with the babies' actual sex no better than a coin toss — kappa 0.0002, indistinguishable from zero. A second study, of births in Ankara, measured the same thing. Nothing observable before a blood test at ten weeks or a scan at eighteen reveals sex.
That is the honest headline, and it belongs first. The accuracy figure usually printed next to this chart — ninety percent and up — has never been reproduced by anyone who measured it. Two studies are the whole of the scientific evidence on this method, and both of them say the same thing.
The Swedish study is the largest of its kind. It covers nearly three million singleton births between 1973 and 2006. For every one of them the researchers reconstructed the mother's lunar age and lunar conception month, then compared the chart's prediction against the birth record.
Accuracy did not improve for younger mothers, older mothers, first-time mothers or any level of education. The Turkish replication, published in 2023, looked at singleton births at one Ankara hospital between 2017 and 2023 and landed in the same place.
None of which makes the chart uninteresting. It is a genuine piece of folk tradition, and it is popular: millions of expecting parents look it up every year. The reason this page exists is that almost every Chinese baby gender predictor serving those families computes the two numbers wrong, and none of them prints what the chart can and cannot predict.
How the Chinese gender chart works
You will see it under several names — the Chinese gender chart, the Chinese gender calendar, the Chinese birth chart, and the Chinese calendar baby gender chart. They are the same grid, and all of them read off the Chinese lunar calendar rather than the one on your wall.
The chart is a grid: the mother's lunar age from 18 to 45 down one side, the lunar month she conceived in from 1 to 12 across the top, and boy or girl in each of the 336 cells. You find your row, find your column, and read the cell. There is no formula underneath and no mechanism — the pattern of boys and girls in the grid does not follow any rule you can state.
Both of those numbers are lunar, and that is where the calculators go wrong. Your lunar age is not your age, and the lunar month you conceived in is not the month on your calendar. Enter the same birth date and date of conception into two Chinese gender predictors and you can get opposite answers, because one of them converted those dates and the other did not.
How to calculate your lunar age, and why “age plus one” is wrong
Traditional East Asian age reckoning starts you at one year old on the day you are born, and adds a year at every lunar new year rather than on your birthday. That is the count the chart was built for.
Most English calculators replace it with a shortcut: take your western age and add one. That agrees with the traditional count for anyone born after the lunar new year. It is wrong by a year for everyone born between 1 January and the new year itself — a window that moves between 21 January and 20 February.
Chinese New Year 1990 fell on 27 January. A woman born on 15 January 1990 is therefore a year older in this count than a woman born on 20 February 1990. The shortcut hands them both the same number, and sends one of them to the wrong row.
| Rule | Where you meet it | What it gets wrong |
|---|---|---|
| Western age + 1 | most English calculators | a year out for anyone born between 1 January and the lunar new year |
| Age + 9 months, +21 if born before 22 February | the German pregnancy sites | pins the new year to a fixed date when the real one moves by a month |
| One at birth, +1 at each lunar new year | the tradition itself, and this page | nothing — but it needs real lunar calendar data to compute |
This calculator does the real thing: it counts the lunar new years between your date of birth and your conception date, using published Hong Kong Observatory conversion tables rather than an approximation.
It shows you both numbers — your ordinary age and your lunar age. When the shortcut would have put you on a different row with a different answer, it tells you what that answer would have been. You should not have to guess which site got it right.
Which lunar month you conceived in
Lunar months begin at a new moon, so they start partway through a Gregorian month and end partway through the next. Roughly the last third of every calendar month belongs to the following lunar month, which means reading your calendar month straight into the chart is wrong about a third of the time.
The tradition also counts from the day of conception, not from the first day of your last period. Most people know their period date or their due date rather than the day itself, so this page will work either out for you — adding two weeks to a period date, subtracting 266 days from a due date — and then shows you the conception date it used, so you can correct it if you know better.
Leap months are the last wrinkle. The lunar calendar inserts a thirteenth month seven times every nineteen years, and the chart has only twelve columns.
The rule usually taught splits the leap month down the middle: conceive in the first half and it counts as the month before, in the second half and it counts as the month after. Two of the Chinese sources we read say plainly that the chart never specified this. So when the two readings give different answers, this page shows you both.
Which Chinese gender chart is the original one?
Every Chinese gender predictor calls its own version the original, usually alongside the same scrap of history: a scroll found in a Qing dynasty royal tomb, kept by the imperial family. No copy of that scroll has ever been shown. In September 2026 we parsed seven published reproductions cell by cell. They were not the same chart.
Four of them agreed on all 336 cells — two English-language sites and two Chinese ones, on unrelated domains. Those four are the chart on this page. The other three deviated: one by 6 cells, one by 36 and one by 38, and the last two did not agree with each other either.
The disagreements cluster in the later lunar months and the upper half of the age range. A conception in the third lunar month gets the same answer everywhere. A conception in the tenth has a real chance of being told the opposite by the next site you open.
We did not majority-vote a single cell. Every one of the 336 on this page is carried by four independent sources with no disagreement between them, and a script in our repository re-checks all seven so the claim can be verified rather than believed.
When a test or a scan can really tell you
Sex is settled at conception by whether the sperm that arrived carried an X or a Y chromosome. Nothing that happens afterwards changes it and nothing observable before about ten weeks reveals it.
A cell-free DNA blood test — NIPT — can read it from about ten weeks. Across 90 studies covering 9,965 pregnancies its sensitivity for fetal sex is around 96.6 percent and its specificity around 98.9 percent. That makes it a screening test, not a diagnosis. The anatomy scan at 18 to 22 weeks usually shows it too, depending on how the baby is lying, and some hospitals decline to look.
Whether a clinic may tell you is a separate question from whether it can. India's PCPNDT Act prohibits disclosing fetal sex outright. Germany allows it only after the twelfth week. Several other countries restrict it.
Until then the Chinese gender predictor is a way to pass an afternoon, which is a perfectly good thing for it to be. If you want the rest of the folklore — the ring test, the heart rate, the Mayan rule — with the same treatment, we have a quiz that walks through all of them.