Nub theory: how accurate it is, by week, from studies

The nub theory sits in an odd place among the ways people guess a baby's sex. Unlike a calendar chart, it rests on something real: the angle has been measured in published studies, and trained sonographers did well with it from about 12 weeks. Unlike a medical test, it is mostly used online, where a parent posts a scan photo and strangers, paid readers or apps give a verdict. This page keeps the two apart.

What is the nub theory?

What is the nub theory in pregnancy? The genital tubercle is a small bump between the legs of a fetus, seen on a scan from side on. Baby nub theory holds that its angle shows whether the baby is a boy or a girl before the sex organs can be told apart directly.

The method behind it comes from research done before invasive tests for sex-linked genetic conditions, where knowing the sex early decided whether a test was needed. Efrat and colleagues examined the genital region on a mid-sagittal view, a side-on slice through the middle of the body, and measured the angle between the genital tubercle and a horizontal line through the skin over the lower spine. They read the fetus as male if the angle was greater than 30 degrees, and as female if the tubercle was parallel or converging, under 30 degrees (Efrat and colleagues, Ultrasound in Obstetrics and Gynecology, 1999).

Nub theory gender: boy and girl nubs, as the theory describes them

The nub theory boy vs girl rule is that single angle:

• A boy, by the theory: the tubercle points upward, at more than 30 degrees to the line along the lower back.
• A girl, by the theory: the tubercle lies flat, parallel to that line or converging with it, at under 30 degrees.

In the 1999 study the angle in male fetuses rose significantly as the fetus grew (Efrat and colleagues, 1999). That is why timing matters so much: early on, a boy's nub has not yet risen, and it can look like a girl's. The nub theory girl vs boy reading is not the shape of the bump or its size, which some online guides describe, but its angle on a correctly aligned side view.

Nub theory ultrasound: what the scan has to show

The studies measured on a true mid-sagittal view, taken for the purpose by an experienced sonographer. Even then, sex could not be assigned in every fetus:

• Efrat and colleagues (2006) could assign sex in 613 of 656 fetuses, 93 in 100, at 12 to 14 weeks (Efrat and colleagues, Ultrasound in Obstetrics and Gynecology, 2006).
• Manzanares and colleagues found it feasible in 608 of 672, 90.5 in 100, at 11 to 13 weeks and 6 days (Manzanares and colleagues, Journal of Clinical Ultrasound, 2016).
• Hsiao and colleagues could determine sex in 441 of 496 (Hsiao and colleagues, 2008).

So under research conditions about 1 fetus in 10 could not be read. A dating scan photo is usually not that view: it is taken to measure the baby and check the dates, and the angle of the probe, the baby's position and the image chosen for printing all change how the tubercle appears.

Nub theory by week

The clearest finding across the studies is that accuracy depends on the week. Gestational age here is counted, as in all pregnancy dating, from the first day of the last period.

Two further studies did not report by single week. Manzanares and colleagues found 532 of 608 readings correct, 87.5 in 100, across 11 weeks to 13 weeks and 6 days, with accuracy rising as the fetus grew and the best results above a crown-rump length of 55.7 mm and a gestational age of 12 weeks and 2 days (Manzanares and colleagues, 2016). Kearin and colleagues, discussed below, looked at routine scans.

Nub theory 11 weeks

At 11 weeks about 7 readings in 10 were right in the two studies that reported that week: 70.3 in 100 (Efrat and colleagues, 1999) and 71.9 in 100 (Hsiao and colleagues, Acta Obstetricia et Gynecologica Scandinavica, 2008). The errors were lopsided. In the 1999 study, boys were wrongly read as girls in 56 in 100 cases at 11 weeks, while only 5 in 100 girls were read as boys (Efrat and colleagues, 1999). A "girl" reading at 11 weeks is therefore the least trustworthy of all.

Nub theory 12 weeks

By 12 weeks the studies found 92 to 98.7 correct readings in 100 (Hsiao and colleagues, 2008; Efrat and colleagues, 1999). The share of boys read as girls fell to 3 in 100 in the 1999 study. Girls were still read less reliably early in the week: 91.5 in 100 at 12 weeks and 0 to 3 days, against 99 in 100 at 12 weeks and 4 to 6 days (Efrat and colleagues, 2006). The authors of the 2006 study advised waiting for a larger fetus before acting on a girl reading for a medical decision. In the UK, the 12-week scan falls in this window, which is why nub theory 12 weeks is the most searched version of the question.

Nub theory 13 weeks

At 13 weeks the studies found 98.3 to 100 correct readings in 100 (Hsiao and colleagues, 2008; Efrat and colleagues, 1999), and the 2006 study found 100 in 100 for girls from 13 weeks (Efrat and colleagues, 2006). Hsiao and colleagues said accuracy at 13 weeks approached that of invasive testing, in their trained hands (Hsiao and colleagues, 2008).

Before 11 weeks

None of the studies above began before 11 weeks. Nub theory at 10 weeks has no published accuracy figure among them, and given how poorly 11 weeks performed, there is no basis for a reading earlier still.

How accurate is the nub theory?

Is nub theory accurate? In the setting it was studied in, yes, from about 12 weeks: trained sonographers, measuring on purpose, on a correct view, did very well. Below 12 weeks it is not reliable, and its mistakes mostly turn boys into girls.

The study closest to everyday scanning is Kearin and colleagues'. Seven sonographers at a private obstetric practice in Australia made predictions during routine scans of 640 fetuses. In the first trimester, 11 to 14 weeks, they were right 75 times in 100 overall, or 91 in 100 when the scans where no prediction could be made were left out; before 12 weeks the overall success rate was 54 in 100. After 14 weeks they were right every time. Results also differed between individual sonographers, and the authors advised discouraging predictions before 12 weeks (Kearin and colleagues, Australasian Journal of Ultrasound in Medicine, 2014).

Nub theory accuracy online is another matter. A parent posting a photo from a 12-week scan, a stranger or a paid reader judging it, an app, or a chatbot asked to read it: none of these has been tested in any study we found. The image was not taken to show the angle, nobody has measured it, and the reader cannot know whether the view is true side-on. Treat a verdict on a photo as a guess, however confident it sounds.

Reliable ways to find out the baby's sex

The sex becomes clearer as the pregnancy goes on. The Office on Women's Health says the external sex organs show by 12 weeks whether the baby is a boy or a girl, and that someone who has an ultrasound in the second trimester or later might be able to find out the baby's sex (Office on Women's Health, "Stages of pregnancy", 2025). In Kearin and colleagues' study, every prediction after 14 weeks was right (Kearin and colleagues, 2014). In the UK, the NHS says that at the 20-week scan, usually done between 18 and 21 weeks, you may be asked if you would like to know the sex of your baby (NHS, "20-week scan", 2024).

A blood test that reads the baby's DNA from the mother's blood, and the ways the NHS and US clinics differ on it, is covered under "Reliable ways to find out the baby's sex" on the Chinese gender predictor page.

Other ways people guess the sex

The nub theory is one of several guessing games. The Chinese gender predictor reads a lunar chart from the mother's age and the month of conception, and a study of 2,840,755 Swedish births found its predictions agreed with the baby's sex no more often than chance (Villamor and colleagues, Paediatric and Perinatal Epidemiology, 2010). The Mayan gender predictor applies an even-and-odd rule to the mother's age and the month of conception. Both are offered on this site as games, with the evidence beside them. The nub is the only one of the three with published accuracy figures, and they hold only for a trained eye on the right view.