# The Chinese Gender Predictor: Fun Tradition or Real Science? _How the lunar chart works, what the studies found, and when you can actually learn your baby's sex._ - Source: https://proba.baby/en/blog/chinese-gender-predictor-explained - Published: 2026-09-05 - Author: ProbaBaby Team - Category: Calculators & Tools - Tags: calculators, first-trimester - Language: English > The Chinese gender chart turns two numbers — lunar age and lunar conception month — into a boy-or-girl answer. Two large studies say it lands at chance. Here is the real math, and the two tests that actually tell you. > **Medical disclaimer.** This article is for general information and does not replace medical advice. The Chinese gender chart is entertainment, not a diagnostic tool. Talk to your midwife or doctor about prenatal screening, and never make a medical decision based on a folk chart. The Chinese gender predictor is a grid that turns two numbers — the mother's lunar age at conception and the lunar month she conceived in — into a single letter: boy or girl. It takes ten seconds, it costs nothing, and it is right about half the time. Two large studies have tested it against real birth records: one covering 2.8 million Swedish births, one covering 17,580 pregnancies. Both found the chart performs no better than a coin toss. That does not make it useless — it makes it a game. Here is how the game works, why it feels uncannily accurate, and what actually tells you. ## What the chart actually is Strip away the styling and the chart is a lookup table. Rows are the mother's lunar age, running 18 to 45. Columns are the lunar month of conception, 1 to 12. Every one of the 336 cells holds a fixed B or G. There is no formula underneath, no weighting, no adjustment for anything about you. You find your row, slide across to your column, and read the letter someone printed there. That matters, because not every version of the chart does the calendar work. Some take your age today and your Gregorian conception month, and either look those up directly or run a parity shortcut on the two numbers instead of reading a cell at all. Those shortcuts can disagree with the real chart, and that is one reason two Chinese gender predictors can hand you two different answers. The 2010 study below took the failure mode seriously enough to model it: the authors simulated "the misuse of the method by failing to convert Gregorian dates into lunar" and reported that it did not change their results. ## The math: lunar age, not birthday age The chart is anchored to the Chinese lunisolar calendar, and both of its inputs are conversions rather than the numbers on your driving licence. ### Nominal age (虚岁) Traditional Chinese nominal age — sui, sometimes written 虚岁 — starts at 1 on the day you are born and adds a year at every lunar new year, not on your birthday. Someone born a month before the lunar new year turns 2 when they are one month old. In practice this runs one to two years ahead of Western age, and the gap depends on where your birthday sits relative to the new year, which itself moves between 21 January and 20 February. The second subtlety: the chart wants your nominal age **at conception**, not today. If you are looking the chart up at 30 weeks pregnant, your age today is often the wrong row. ### The lunar month The lunar month is not your Gregorian month either. Chinese months begin at the new moon, so a conception on 3 February might land in the twelfth month of the previous lunar year. And roughly every three years the calendar inserts a leap month — an intercalary repeat of an existing month — to keep it aligned with the solar year. The traditional chart has no row for a leap month, so implementations have to decide what to do with one. The usual convention is to fold it into the number of the month it repeats. Then there is the input itself. Most people do not know their conception date; they know when their last period started. If you are estimating conception the way [due date calculators estimate your conception date](/blog/how-due-date-calculators-work), you are feeding an approximation into a lookup table that is sensitive to which side of a new moon you land on. The 2023 study below ran into exactly this — it had to assume conception was the first day of the last period plus 14 days, and lists that assumption as a limitation of its own. ## Where the chart came from The story usually told online is that the chart is around 700 years old, was discovered in a royal tomb near Beijing, and that the original sits in a scientific institute in the city. The version that reached the peer-reviewed literature is narrower: the 2023 *Actual Gynecology and Obstetrics* paper repeats the legend as saying the chart was invented at the beginning of the Qing dynasty (1644-1912) with the help of the I Ching, and was found in an emperor's tomb "in the last century". That is under 400 years, not 700 — and it is a legend the paper is reporting, not a finding it verified. Nobody has produced peer-reviewed documentation of the original document, its dating, or its custody. Published copies do not entirely agree with each other either: when ProbaBaby transcribed two published charts and compared them cell by cell to build its own table, the rows for ages 19 to 45 matched exactly and the age-18 row did not. Treat the provenance the way you would treat any piece of folklore: culturally real, historically fuzzy, and not evidence of anything. ## What the studies found This is the part that settles it. In 2010, Villamor and colleagues tested the chart against Swedish population birth data: 2,840,755 singleton births between 1973 and 2006. They estimated each mother's lunar age and lunar conception month, ran a published version of the method, and compared the prediction to the baby's recorded sex. The agreement statistic (kappa) came out at 0.0002, with a confidence interval of −0.0009 to 0.0014 straddling zero — where 0 means pure chance and 1 means perfect agreement. Accuracy was not modified by year of conception, maternal age, education, body mass index, or parity. They also re-ran a 1,000-birth subset using a website's own lunar-date conversion, and kappa there was −0.02. Their published conclusion is that the method is "no better at predicting the sex of a baby than tossing a coin." A 2023 study repeated the exercise on a different population: 17,580 singleton pregnancies delivered at one hospital in Ankara, Türkiye, between 2017 and 2023, with twins, IVF pregnancies, uncertain last-period dates and ambiguous genitalia at birth all excluded. Kappa came out at 0.002, p = 0.797. Same answer, different country, different decade. That paper also points at a third dataset in its discussion: a 2009 study by Katz and colleagues of singleton births, 38,394 patients, in which the chart's predictivity was reported at 50.2%. Two datasets tested directly, roughly 2.86 million pregnancies between them, and the chart lands exactly where chance predicts. ## Why it still feels accurate If the chart is a coin toss, why does everyone know someone it "got right"? Because a coin toss is right half the time, and you only hear about the hits. Three friends use the chart; roughly one and a half of them get the correct answer; those are the ones who mention it at the baby shower. Nobody posts "the chart said girl and we had a boy." Two more things push it along. People often check the chart *after* a scan and remember it as having predicted correctly. And when the chart and the scan disagree, most people quietly drop the chart — it never enters the tally as a miss. That is confirmation bias doing exactly what it does everywhere else. ## The other folk methods The chart has company: the [Mayan gender prediction method](/tools/mayan-gender-predictor), the wedding-ring-on-a-string test, carrying high versus low, sweet cravings versus salty, and the fetal-heart-rate rule (above 140 beats per minute means girl). None of them has evidence behind it, and the one that has been measured most carefully failed outright. The heart-rate one is the most persistent, and it has been measured directly. In 477 singleton pregnancies scanned before 14 weeks, the average rate was 151.7 bpm for female fetuses and 154.9 bpm for male ones — a gap of about 3 bpm that did not reach statistical significance (p = 0.13). The rule has no signal to stand on. Folk methods persist because they are fun, shareable, and cost nothing to be wrong about. That is a fine reason to keep them, as long as the receipt stays with the gift. ## When you can actually know Two things genuinely tell you, and both are part of routine [prenatal testing during pregnancy](/medical-tests). **Cell-free DNA screening (NIPT), from around 10 weeks.** A blood draw from the mother picks up placental DNA circulating in her bloodstream. The test needs the fetal fraction — the share of that DNA coming from the placenta — to be above about 4%, which MedlinePlus says typically happens around the tenth week; below that, the lab may not be able to run it or may return a false negative. Alongside its main job of screening for trisomy 21, 18 and 13, NIPT looks at the X and Y chromosomes. One caution worth keeping in view: MedlinePlus is explicit that NIPT "is a screening test, which means that it will not give a definitive answer". **Ultrasound, more reliable as pregnancy goes on.** A 2025 systematic review and meta-analysis included 28 studies; the nine reporting first-trimester accuracy put it at a mean of 85% and a median of 87%, and it climbs sharply within a few weeks — sensitivity was 69% at 11 weeks, 89% at 12 and 96% at 13. For the second and third trimesters the review reports mean and median rates of 92% and 99%. In the UK, the 20-week screening scan is usually done between 18 and 21 weeks; it looks for 11 specific conditions and sex is not one of them, though the NHS says you may be asked whether you would like to know. So the honest timeline is: nothing before about 10 weeks, a blood test that can report the sex chromosomes from around then, and a scan that gets steadily more reliable through the end of the first trimester and into the second. ## How ProbaBaby helps In the ProbaBaby app, the Chinese Gender Predictor does the calendar work rather than shortcutting it. It takes two dates — the mother's birth date and the conception date — converts both with the Chinese calendar, and works out the traditional nominal age (虚岁) **at conception**, not the age today. It then reads the answer out of a transcribed chart covering ages 18 to 45 and months 1 to 12. A conception falling in an intercalary month is folded into the number of the month it repeats, because the traditional chart has no leap row. If the nominal age lands outside 18 to 45, there is no cell to read, and the app says so rather than extrapolating one. Before the reveal, a live preview shows the lunar age and month it resolved from your dates, and the chart itself is drawn on screen with your cell marked, so you can see exactly where the answer came from. The result carries the app's own disclaimer, which does not hedge: "The Chinese Gender Predictor is based on an ancient chart and is for entertainment purposes only. It has no scientific basis and should not be used for medical decisions. The actual accuracy is approximately 50% - the same as random chance." There is also a [Chinese gender predictor](/tools/chinese-gender-predictor) here on the website. ## FAQ **How accurate is the Chinese gender predictor?** About 50% — the same as a coin toss. A study of 2,840,755 Swedish births found a kappa agreement statistic of 0.0002 between the chart's prediction and the baby's actual sex, and a 2023 study of 17,580 singleton pregnancies in Türkiye found kappa 0.002 with p = 0.797. **What two numbers does the Chinese gender chart use?** The mother's lunar (nominal) age at conception and the Chinese lunar month in which she conceived. Both are calendar conversions, not the birthday age and Gregorian month many online versions use. **What is lunar age, and why is it different from my real age?** Traditional Chinese nominal age counts you as 1 at birth and adds a year at each lunar new year, not on your birthday. So it usually runs one to two years ahead of your Western age, depending on where your birthday falls. **Is the Chinese gender calendar really 700 years old?** There is no verified provenance. The version of the legend repeated in the peer-reviewed literature puts the chart at the start of the Qing dynasty (1644-1912) and says it was later found in an emperor's tomb — under 400 years, not 700. Either way that is a legend a paper reports, not a documented history. **When can I actually find out my baby's sex?** Cell-free DNA screening can report the sex chromosomes once the fetal fraction is above about 4%, which typically happens around the tenth week. Ultrasound gets steadily better: one 2025 meta-analysis reports pooled sensitivity rising from 69% at 11 weeks to 89% at 12 and 96% at 13, with higher accuracy in the second and third trimesters. In the UK the 18-to-21-week scan checks for 11 specific conditions, and sex is not one of them, though you may be asked whether you want to know. **Does the chart work for twins?** No. The chart holds one letter per cell, so it returns one answer for one conception and cannot say anything meaningful about two babies who may not share a sex. Treat it as a single coin toss printed on a grid. **Is it harmful to use the chart?** Not if you treat it as a game. It becomes a problem when it drives real decisions — buying clothes, painting a nursery, or announcing a sex before a screening test or scan has indicated it. ## Sources - Villamor E, Dekker L, Svensson T, Cnattingius S. Accuracy of the Chinese lunar calendar method to predict a baby's sex: a population-based study. Paediatr Perinat Epidemiol. 2010;24(4):398-400 — 2,840,755 singleton Swedish births, 1973-2006; kappa 0.0002 [95% CI -0.0009, 0.0014]: https://pubmed.ncbi.nlm.nih.gov/20618730/ - Jafarzade A, Aghayeva S, Mungan TM, Biri A, Ekiz OU. Predicting Baby Gender Using the Chinese Lunar Calendar. Actual Gyn. 2023;15:57-60 — 17,580 singleton pregnancies, Ankara 2017-2023; kappa 0.002, p = 0.797: https://www.actualgyn.com/en/article/2023/281 - MedlinePlus Genetics (NIH): What is noninvasive prenatal testing (NIPT)? — Fetal fraction must be above 4%, typically around the tenth week; NIPT is a screening test: https://medlineplus.gov/genetics/understanding/testing/nipt/ - NHS: 20-week screening scan — Usually done between 18 and 21 weeks; screens for 11 conditions, sex not among them: https://www.nhs.uk/pregnancy/your-pregnancy-care/20-week-scan/ - Amankona E, Donkor A, Agyei BA, et al. Effectiveness of prenatal ultrasound in fetal sex identification: a systematic review and meta-analysis. J Ultrasound. 2025;28(3):563-574 — 28 studies; first trimester mean 85% / median 87%; later trimesters mean 92% / median 99%: https://pubmed.ncbi.nlm.nih.gov/39798051/ - McKenna DS, Ventolini G, Neiger R, Downing C. Gender-related differences in fetal heart rate during first trimester. Fetal Diagn Ther. 2006;21(1):144-7 — 477 singleton pregnancies; female 151.7 bpm vs male 154.9 bpm, p = 0.13: https://pubmed.ncbi.nlm.nih.gov/16354993/ --- ProbaBaby (listed on the App Store as “Pregnancy & Baby Tracker - PB”) is a free iPhone, iPad and Apple Watch app for pregnancy, newborn and postpartum tracking, built for twins, in 32 in-app languages, with an optional AI baby-face prediction that computes Mendelian genetics (Punnett squares) from the traits you enter before any image is generated. Tracking, tools, PDF reports and family sharing are free; only AI predictions use credits — 5 starter credits cover your first Scientific prediction — and no subscription is required. Requires iOS 26 or later. Not available on Android.