Turn Your Love Story Into a Glimpse of Tomorrow With Our AI Baby Generator
One-time payment, no subscription
Partly. Nub theory is the only method on this page with published accuracy data behind it. In the largest study we could find, it was right 72% of the time at 11 weeks and 88% at 13 — better than a coin flip, well short of the 95–100% figures repeated online, and not good enough that the researchers would use it clinically.
Every fetus grows a genital tubercle at around nine weeks, and for a while it looks the same whatever the chromosomes say. Between roughly 11 and 14 weeks it starts to differentiate: in male fetuses it angles upward away from the spine, and in female fetuses it stays closer to parallel. Nub theory is the practice of reading that angle off a mid-sagittal ultrasound image.
This is why the method has a window rather than a date. Before about 11 weeks there is nothing to distinguish. After about 14, the anatomy has developed far enough that a sonographer is looking at the genitals themselves rather than inferring from an angle.
The most useful study we found measured genital tubercle angles in 455 fetuses and used a 42-degree threshold. Correct prediction ran at about 72% between 11 and 11 weeks 6 days, 86% between 12 and 12 weeks 6 days, and 88% between 13 and 13 weeks 6 days.
The authors' own conclusion is the part worth keeping: accuracy was high from 12 weeks onward, but still not high enough to use clinically in pregnancies at risk of a serious X-linked condition. A method that researchers will not lean on for a decision that matters is a method to enjoy rather than plan around.
Popular articles quote 90%, 98%, sometimes 100% at 12 or 13 weeks. Those numbers rarely carry a citation, and where the underlying method is described, the reason for the gap is usually that the count only includes scans where the nub was clearly visible and confidently read.
That is a selection effect, not an accuracy rate. Throwing out the unclear images and scoring only the easy ones will raise any method's hit rate. The honest figure includes the scans where someone looked, could not tell, and guessed anyway — which is most of what happens in practice.
There is also no agreement on where the line sits. The published study used a 42-degree cutoff; the guides circulating online generally use 30 degrees. Those two thresholds disagree about a real range of real images, so two people can read the same nub correctly by their own rule and reach opposite answers.
A nub reading needs a clean mid-sagittal view, with the fetus lying in a position that shows the tubercle in profile and the spine in the same plane. Plenty of 12-week scans never produce that image, because babies move and sonographers are looking for other things.
Experienced readers also report over-calling boys, because an angle exaggerated by a slightly oblique view looks male more often than it looks female. If a guess matters to you, treat a nub reading as a strong hint at 13 weeks and little more than a lean at 11.
The two methods that do give an answer are a non-invasive prenatal test, which reads placental DNA from a sample of the mother's blood from about nine or ten weeks, and the anatomy scan at eighteen to twenty-two weeks.
About 86% in the largest study we found, which measured 455 fetuses using a 42-degree threshold. That is meaningfully better than guessing, and meaningfully worse than the 98% figures that circulate without a source.
Barely. The same study put accuracy at 72% between 11 and 11 weeks 6 days. The tubercle has only just begun to differentiate, so the angle you are reading is small and easily distorted by the fetus's position.
There is no settled answer, which is itself a problem with the method. The published study used 42 degrees; the guides shared in pregnancy forums generally use 30. The same image can be read either way depending on which rule the reader follows.
The common explanation is geometry. If the scan is even slightly off a true sagittal plane, the apparent angle of the tubercle is exaggerated, and an exaggerated angle reads as male. Readers describe over-predicting boys rather than girls as a result.
Not as a basis for decisions. The researchers behind the accuracy figures above concluded explicitly that it was not reliable enough for pregnancies at risk of a serious X-linked condition, which is exactly the situation where an early answer would be most valuable.
A non-invasive prenatal test from around nine or ten weeks, which reads placental DNA in the mother's blood, or the anatomy scan at eighteen to twenty-two weeks. Both are far more reliable than any angle measured at twelve weeks.