Does nub theory actually work? What the research says
Search "nub theory accuracy" and you'll see everything from "50-50, basically a guess" to "99.2% accurate, scientifically proven." Both of those are a bit misleading. Here's what the actual research says.
The real studies behind it
Nub theory isn't something the internet invented from nothing. Sonographers first wrote it up formally in a 1999 study in the journal Ultrasound in Obstetrics & Gynecology, which looked at the genital tubercle's angle against the spine in around 170 pregnancies and found accuracy climbed sharply with gestational age, from roughly seven in ten correct at 11 weeks to close to certain by 13 weeks. A follow-up study in 2001 found broadly similar patterns. More recently, a 2024 study measuring both the tubercle angle and a second anatomical distance in over 300 pregnancies again found the angle to be a genuinely useful, statistically significant signal, not just internet folklore.
Why online accuracy claims are all over the place
You'll see services advertise a single flat number, "99% accurate" is a popular one, without mentioning that accuracy in the actual research depends heavily on how far along the pregnancy is and how clearly the nub can even be seen. Earlier studies also don't always agree with each other: a couple of larger, later studies found meaningfully lower accuracy than the original 1999 paper, particularly at earlier gestational ages. Quoting one flattering number from one favourable study, without the caveats, makes for a better advert than an honest range.
Where it can still go wrong
Even in the research, the nub simply isn't readable on every scan. Before around 11 and a half weeks, researchers could only get a usable angle roughly four times in ten, rising steadily as the pregnancy progresses. Baby's position, the angle of the probe, and general image clarity all affect whether the angle can be judged with any confidence at all, which is exactly why a blurry or awkwardly angled photo pulls your score down rather than getting forced into a guess.
How this shapes what Kay does
This is why Kay weighs gestational age and image clarity into her reading rather than applying one flat percentage to every scan, and why a photo that's hard to read honestly brings your confidence score down instead of being rounded up. It's also why, if Kay can't reach 75% confidence, you're told that plainly and refunded, rather than being handed a guess dressed up as certainty.