Two NASA probes built in the early 1970s spent three decades drifting slightly off course, decelerating by about 8.74 x 10⁻¹⁰ metres per second squared more than the Sun’s gravity could account for, a force roughly ten billion times weaker than the pull you feel standing on a sidewalk. The discrepancy was small enough to be invisible on any practical scale and large enough to threaten the foundations of physics. According to Astronomy Magazine, the effect kept the Pioneers thousands of kilometres closer to Earth, every year, than the equations said they should have been.
For thirty years it had a name. The Pioneer anomaly.
The probes that produced it, Pioneer 10 and Pioneer 11, were launched in 1972 and 1973 to fly past Jupiter and Saturn and then keep going forever. They were the first human objects aimed beyond the Solar System. They also happened to be, by accident of engineering, the most precisely tracked spacecraft ever flown.
Why the Pioneers could be measured so exactly
Most deep-space probes are three-axis stabilised. They fire small thrusters constantly to keep their antennas pointed at Earth, and every one of those puffs is a tiny shove that smears any subtle gravitational signal you might be hoping to detect.
The Pioneers were different. They were spin-stabilised, rotating around a central axis like a thrown frisbee, and they needed almost no thruster firings to stay on course. Combined with the Doppler tracking system at NASA’s Deep Space Network, this allowed engineers to measure the probes’ acceleration to about 10⁻¹⁰ metres per second squared. That is the kind of precision where the gravitational tug of a distant comet field starts to matter.
The Voyagers, often confused with the Pioneers in popular memory, cannot be tracked anywhere near this finely. They use their thrusters too often. The Pioneers, freely coasting, were a uniquely sensitive instrument for measuring the gentle physics of the outer Solar System.
That sensitivity is what produced the mystery.
What John Anderson noticed in 1980
By 1980, Pioneer 10 had passed roughly 20 astronomical units from the Sun, somewhere out past Uranus. John Anderson at NASA’s Jet Propulsion Laboratory was building an algorithm to predict its motion exactly. He included the gravity of every planet, the pressure of sunlight, the delay of radio waves, the rotation of the Earth, the drag of solar wind, the gravitational influence of the asteroid belt.
The model still came up short. The probe was slowing down a little more than predicted, and the extra deceleration pointed back toward the Sun. The same pattern would later appear in Pioneer 11’s tracking data as it travelled further out from Saturn.
It was not a large effect. As Popular Science later detailed, the constant sunward pull worked out to about a ten-billionth of Earth’s surface gravity. But it did not go away. Year after year, the Pioneers fell a few hundred miles further behind their predicted positions. Both of them. By the same amount.
Anderson’s first thought was that fuel was leaking from the thrusters. That hypothesis aged poorly. The probes had stopped using their propulsion systems years earlier, and the anomaly stubbornly continued.
A coincidence that almost broke physics
In 1994, Anderson got an email from Michael Martin Nieto, a cosmologist at Los Alamos National Laboratory. Nieto was interested in alternatives to Newton’s inverse-square law for gravity, including a theory called modified Newtonian dynamics. He wanted to know how confident NASA was in its gravitational measurements at large distances.
Anderson told him about the Pioneer numbers. When Nieto read the exact value of the anomaly, he later said, he rocked back so hard in his office chair that the wheels rolled.
The value of the anomalous acceleration was almost identical to the speed of light multiplied by the Hubble constant, the so-called cosmic acceleration that describes the expansion of the universe. If the match was real, the Pioneers were measuring something fundamental about cosmology while drifting past Pluto.
In 1998, Anderson, Nieto, Slava Turyshev and others published their findings in Physical Review Letters. The same year, astronomers announced the discovery of dark energy. The two stories collided in the imagination of theoretical physicists.
What followed was, by any measure, a small academic stampede.
Thirty years of papers, conferences, and exotic theories
By the late 2000s, the Pioneer anomaly had inspired close to a thousand academic papers. Researchers proposed dark matter halos around the Solar System, modifications to general relativity, a local blueshift region surrounding the Sun, and the idea that the anomaly was a signature of cosmological time itself accelerating.
One proposal documented by IFLScience argued that the anomaly was not a force acting on the probes at all, but a manifestation of universal expansion changing the background gravitational potential, which would shift the radio frequencies as the signals travelled.
Another team, working on a framework they called conformal cosmology, suggested the Pioneers were passing through a local region where light was subtly blueshifted, affecting the Doppler measurements that defined the anomaly in the first place.
The question on every cosmology desk was the same. Had we got gravity all wrong?
Critics pushed back. If modified Newtonian dynamics or dark matter were responsible, the outer planets should be drifting in the same way, and they were not. The most likely explanation, the sceptics said, was heat. The plutonium-238 inside each Pioneer’s radioisotope thermoelectric generators produced roughly 2,500 watts at launch. Most of that energy was not converted to electricity. It radiated away as warmth.
If the warmth came off the spacecraft asymmetrically, even slightly, the recoil could push the probes backward. A five percent imbalance in thermal radiation would be enough to produce the entire anomaly.
Why the heat theory took so long to win
The heat hypothesis was on the table from the beginning, but the Pioneer team initially ruled it out. The RTGs were mounted on long booms, far from the body of the probe, and only a small fraction of their radiation should have hit the spacecraft. The decay of plutonium-238 also meant the heat output dropped over time, while the anomaly appeared constant.
Both of those objections turned out to be wrong, or at least incomplete.
By the late 2000s, Slava Turyshev at JPL and Viktor Toth, working in Canada, had begun an extraordinary archival project. The original Doppler tracking records were scattered across paper printouts and 7- and 9-track magnetic tapes, some literally rescued from boxes under stairwells at JPL on their way to being thrown out. The Planetary Society sent out appeals to its members to help fund the data recovery effort, and old hardware had to be located and revived to read the tapes and convert decades of telemetry into modern formats.
With the full dataset finally in hand, Turyshev and Toth built a detailed thermal model of each probe. They included heat from the RTGs reflecting off the back of the high-gain antenna, heat from the electronics compartment radiating in the direction of motion, and the warming effect of sunlight on the trailing face of the spacecraft.
The answer was the probes themselves
The combined picture showed something the earlier analyses had missed. Heat from the RTGs was bouncing off the back of the dish-shaped antenna and being directed forward, in the direction the probes were travelling. Heat from the electronics box was radiating the same way. The probes were, in effect, gently shining infrared light out their front end, and the recoil from those photons was pushing them backward toward the Sun.
Once Turyshev’s team included these thermal recoil forces in the trajectory model, the anomaly disappeared. Their analysis also showed that the apparent constancy of the anomaly had been an artifact of limited data. Over longer timescales, the effect was slowly decreasing, consistent with the radioactive decay of the plutonium-238 fuel.
The probes were not telling us anything new about gravity. They were warming the inside of their own dish antennas and being pushed by their own infrared glow.
The force involved is almost laughably small. A billionth of Earth’s gravity is the kind of pressure you would get from the weight of a few grains of pollen resting on your palm. Across decades and billions of kilometres, it added up to thousands of kilometres of cumulative drift.
What the resolution did and did not change
Einstein and Newton were not dethroned. The cosmic acceleration coincidence that Nieto had noticed remained a coincidence. As subsequent atomic clock experiments have continued to demonstrate, general relativity passes every precision test thrown at it.
As La Brújula Verde noted in a 2024 retrospective, the resolution closed a chapter that had occupied entire scientific careers. Other anomalies remain. The flyby anomaly, in which probes like Galileo, NEAR, Cassini and Rosetta have picked up tiny unexpected velocity changes during Earth gravity assists, has not been fully explained.
The Pioneers themselves are now silent. Pioneer 10 last contacted Earth in January 2003 from a distance of about 80 astronomical units. Pioneer 11 went quiet in 1995. Both are still out there, drifting through the Milky Way, carrying the gold plaques that Carl Sagan helped design in three weeks, alongside Frank Drake and Linda Salzman Sagan, in case anyone, ever, finds them.
The lesson buried in the data tapes
The Pioneer story is sometimes told as a cautionary tale, a reminder that exotic explanations should wait until the boring ones have been ruled out. That framing is partly right and partly unfair. The original team did consider heat in the 1980s and 1990s. They ruled it out based on the best modelling available at the time. The fault was not enthusiasm for new physics. The fault was that the full thermal picture required data that had been left in cardboard boxes.
The deeper point is what the Pioneers proved is possible. Two probes built before the first home computer, carrying instruments designed in the late 1960s, allowed engineers on Earth to measure their motion to a precision that revealed the gentle pressure of their own infrared warmth across forty years of flight. The people who eventually solved the puzzle spent years hunting magnetic tapes and writing emulators for hardware that no longer existed.
That is a strange kind of scientific labour. Half archaeology, half astrophysics. The Pioneers were sending signals from beyond Neptune. The signals were arriving. The question was whether anyone still had the machines to read them.
The probes are now well past 100 astronomical units from the Sun, beyond the reach of any antenna built so far. Each one is still being pushed backward, very gently, by the last warmth of decaying plutonium they cannot feel and we can no longer hear.
