For planetary scientists, it was the boldest claim in a generation: an unseen extra planet, as much as 10 times the mass of Earth, lurking on the Solar System’s frontier, beyond Neptune. But the claim looks increasingly shaky, after a team of astronomers reported last week that the orbits of a handful of distant lumps of rock are not bunched together by the gravity of Planet Nine, as its proponents believe, but only seem clustered because that’s where telescopes happened to be looking.

Planet Nine supporters aren’t backing down yet, but one skeptic not involved with the new work says she is “very happy” to see it. The study has carried out “a more uniform analysis” than done previously of the far-off rocky bodies known as known as Trans-Neptunian objects (TNOs), says astronomer Samantha Lawler of the University of Regina, who has tried and failed to simulate the clustered orbits in computer models with an extra planet.

Mike Brown and Konstantin Batygin of the California Institute of Technology made headlines worldwide in 2016 with their prediction for a distant Planet Nine. They based their conclusion on a study of six TNOs, each smaller than Pluto, in extremely elongated and tilted orbits around the Sun. The orbits of these “extreme” TNOs were bunched together, Brown and Batygin said, because Planet Nine’s gravity had nudged them there over billions of years. Several more extreme TNOs discovered since then seemed to cluster as well. “I would argue that the relevant [Planet Nine] data set is in pretty good shape,” Batygin says.

Lawler and other astronomers were concerned about selection biases, however. Given how small and dark extreme TNOs are, they are only visible—if at all—during their closest approach to the inner Solar System, and often only if they are not observed against the bright backdrop of the Milky Way’s disk. Critics of the Planet Nine claim said the apparent clustering of the discovered TNOs might only be because that’s where telescopes were looking or were most sensitive. “Every survey has biases,” Lawler says. “Some are aware of them, some are not.”

Planet Nine is claimed to be five to 10 times as massive as Earth, in an orbit well beyond Neptune.

A team led by Kevin Napier of the University of Michigan, Ann Arbor, decided to test whether selection bias was playing a role. They gathered 14 similarly distant TNOs discovered by three different surveys: the Dark Energy Survey (DES), which uses the Blanco Telescope in Chile; the Outer Solar System Origins Survey on the Canada-France-Hawaii Telescope in Hawaii; and a third that used a variety of telescopes. All three had well characterized selection biases. None of the 14 TNOs was among the original six invoked by Brown and Batygin.

• 3 proofs that debunk flat-Earth theory | NASA's Michelle Thaller
• 'You're an idiot': Nickson Clark explodes at flat-Earther on Mai FM
• 5 Flat Earth Arguments DEBUNKED
• Aliens made agreement with US President Donald Trump to hide on Earth - former Israeli space head

Napier says the team took account of when and where the telescopes pointed, and how sensitive they were to faint objects. With those data, the team calculated a “selection function” that varies across the sky. And sure enough, the extreme TNOs found by all three surveys were in or near areas where selection function was highest, the team reported on 11 February in a paper posted to the arXiv and accepted by the Planetary Science Journal. As a result, Napier says, the team could not reject the null hypothesis that the extreme TNOs are uniformly distributed around the Solar System, which would rob Planet Nine of its foundational evidence. The clustering “is a consequence of where we look and when we look,” he says. “There’s no need for another model to fit the data.”

Batygin doesn’t accept that conclusion. He points out that the DES survey looked largely in the area of the sky where the TNO cluster he and Brown identified resides and found more extreme TNOs. So ruling out clustering is “not logical,” he says. “The more relevant question to ask is: Can their analysis distinguish between a clustered and uniform distribution, and the answer appears to be no,” he says.

Napier acknowledges that trying to draw conclusions from a sample of 14 TNOs is tricky. “There’s only so much statistical power you can draw with so few objects,” he says. The matter is unlikely to be settled, he adds, until the Vera C. Rubin Observatory—a powerful new survey telescope being built in Chile—starts to observe in 2023. Its survey will have well-defined selection biases and is likely to detect hundreds of new extreme TNOs. That, Napier says, “will be like Christmas morning.”