The Three-Body Problem Explained: Why Astronomy's Simplest Question Has No Answer

Take three objects. Specify their masses. Specify, with infinite precision, where they are and how fast they are moving at one moment in time. They attract each other gravitationally. Predict where they will be in a thousand years.

This is the three-body problem. It is the simplest gravitational question that goes beyond what Newton's mechanics can answer in a closed form. It has no general analytic solution. It is, in the technical sense of the word, chaotic. It has been an open problem for more than three centuries.

It is also one of the most beautiful problems in science, because the impossibility of a clean answer turns out to reveal something deep about how the physical world works. Chaos, as the mathematicians of the 20th century clarified, is not noise. It is structure we cannot predict.

What "two-body" gets you, and what it doesn't

For two objects — say, the Earth and the Sun — Newton's laws give you a clean, closed-form answer. The orbit is a conic section: an ellipse, a parabola, or a hyperbola, depending on the energy. Kepler's three laws describe it. You can write the position of either body at any future time as a precise mathematical function of the initial conditions. The math is exact. The prediction is, in principle, perfect for all time.

This is what made Newton's Principia in 1687 such a stunning achievement. The motion of the planets — observed but unexplained for thousands of years — turned out to be the consequence of one inverse-square force law, computable from two pieces of information per body.

The natural next question was: what about three?

Where it breaks

Add a third body and the equations of motion still look clean. Each body experiences the gravitational pull of the other two, summed as vectors. Three bodies gives you a coupled system of nine second-order differential equations (three for each body), or six if you exploit conservation of momentum and centre of mass.

The equations are simple. The integrals are not. There is no general way to write the future positions of three gravitating bodies as a closed mathematical expression of the initial conditions. The system can be integrated numerically — given a computer and a step size — but no formula exists that, plugged into the initial conditions, spits out where everyone will be at time T.

This was understood implicitly by the late 1700s and proven definitively by Henri Poincaré in 1889.

Poincaré's contribution

In the 1880s, King Oscar II of Sweden offered a prize for the solution to the n-body problem in celestial mechanics. Poincaré attacked the three-body case. His prize-winning paper was, on submission, generally considered correct.

Then, while preparing the paper for publication, Poincaré realised he had made a mistake. The corrected version showed something far more interesting than what he originally claimed. Not only had he failed to find a general solution to the three-body problem — he had proven, structurally, that no such solution could exist, at least not in the kind of well-behaved form that the two-body problem produced.

The reason was what we now call sensitive dependence on initial conditions. In a stable two-body system, a small change to the starting position of either body produces a small change to the future orbit. The system is predictable in the everyday sense. In the three-body problem, certain configurations have the opposite property: an arbitrarily small change in starting conditions produces wildly different future trajectories.

This is the technical definition of chaos. The system is fully deterministic — the equations are exact, there is no randomness — but its long-term behaviour is, in practice, unpredictable, because we can never know the initial conditions to infinite precision.

Poincaré's work, mostly ignored at the time, became the seed of 20th-century chaos theory. Edward Lorenz's discovery of strange attractors in atmospheric models in the 1960s, Mitchell Feigenbaum's universal constants in the 1970s, the explosion of dynamical systems research that followed — all of it traces back to the three-body problem and the moment Poincaré realised what he'd actually found.

Numerical simulation: the modern workaround

If you can't solve the three-body problem analytically, you can simulate it. Specify the masses and initial conditions. Choose a small time step. Update each body's position and velocity using Newton's laws and the current positions of the other two. Repeat for as many steps as you want.

Modern numerical integrators — symplectic methods that conserve energy over long simulations, adaptive step-size methods that get smaller when bodies pass close to each other — have made it possible to model gravitational systems with extraordinary precision. NASA does this routinely for spacecraft trajectory planning. Astronomers do it for galactic dynamics, asteroid tracking, and exoplanet orbital stability studies.

The catch is that numerical simulation is not a true solution. Two simulations of the same physical system, started with initial conditions that differ in the tenth decimal place, will diverge over long enough timescales. For chaotic configurations, the divergence happens fast. There is a mathematical horizon, called the Lyapunov time, beyond which prediction is effectively impossible no matter how good your computer is.

For the inner solar system, the Lyapunov time is roughly five million years. We can predict planetary positions a few million years out with reasonable confidence. We cannot predict, with any precision, where Mercury will be in a billion years.

The restricted three-body problem and Lagrange points

There is one famous special case that does have an elegant partial solution. The restricted three-body problem assumes one of the three bodies is so much smaller than the other two that its mass can be neglected in the gravitational calculation. The two large bodies orbit each other in a clean two-body solution, and the small body moves in their combined gravitational field.

This is a good model for, say, a spacecraft moving through the Earth-Sun or Earth-Moon system. In this restricted case, Joseph-Louis Lagrange showed in the 1770s that there are five points in the rotating frame of the two large bodies where the gravitational forces and centrifugal force balance exactly. A small object placed at one of these Lagrange points will, in principle, stay there.

L1 and L2 are unstable equilibria — like a ball balanced on top of a hill — but unstable in a useful way: a spacecraft can hold position at L1 or L2 with very modest fuel consumption. L4 and L5 are genuinely stable, which is why they accumulate Trojan asteroids in the Sun-Jupiter system.

This is why the James Webb Space Telescope sits at Sun-Earth L2, roughly 1.5 million kilometres beyond Earth's orbit. The L2 position keeps JWST in line with Earth and Sun, which lets it use a single shield to block heat from both at once. The telescope makes small thrust corrections every few weeks to maintain its orbit around the unstable equilibrium point. Solving this kind of orbital mechanics is one of the practical achievements of three-body theory: even though we cannot solve the general problem, we can engineer specific solutions for specific configurations.

The 2023 family of new periodic solutions

The three-body problem mostly produces chaotic trajectories, but there are special configurations — periodic orbits — where three bodies return to the same configuration after some fixed time. These are mathematically rare and physically delicate, but they exist. The most famous is the figure-eight orbit, discovered by Christopher Moore in 1993, in which three equal masses chase each other around a single figure-eight curve.

In 2023, a team of researchers used massive computer searches to discover a new family of periodic three-body solutions — hundreds of them, in configurations that no one had previously catalogued. Some involve bodies of different masses. Some have orbits that look like complex knots in 3D space. Almost all of them are unstable: a tiny perturbation would send the system tumbling into chaos. But they are, mathematically, real solutions.

The fact that we are still finding new periodic three-body orbits, more than 300 years after the problem was formulated, is a reasonable indication of how rich the mathematical structure underneath the chaos actually is.

Cixin Liu's novel, briefly

Cixin Liu's 2008 novel The Three-Body Problem (translated into English in 2014, adapted by Netflix in 2024) takes the physical impossibility of the problem and makes it a plot device. An alien civilisation lives on a planet with three suns, and the unpredictable gravitational chaos of the three-star system is the existential threat that drives them to invade Earth.

The astrophysics is loose. Real triple-star systems exist — Alpha Centauri is one — but in stable configurations where one component orbits a tight binary at a great distance, which keeps the system effectively two-body for long timescales. A planet orbiting in the chaotic regime Liu describes would more likely be ejected from the system than survive long enough to evolve a civilisation.

The novel works anyway because the underlying instinct is correct: the three-body problem is the moment where deterministic physics starts producing unpredictable outcomes, and that instinct — that the universe contains regions where prediction simply fails — is one of the most unsettling truths physics has produced.

What chaos actually means

The deepest lesson of the three-body problem is the one that took the longest to settle. For two centuries after Newton, scientists assumed that the universe was deterministic in the strong sense: given enough computing power and enough precision, any future state could be predicted from any present state. Pierre-Simon Laplace's "demon" — the imagined intelligence that, knowing the position and velocity of every particle, could predict everything — was the dream of classical physics.

The three-body problem broke that dream. The universe is deterministic in the equations and unpredictable in the outcomes. Determinism and predictability, which philosophers had assumed were the same thing, turn out to be different.

Chaos is not noise. Noise would mean the system is fundamentally random. Chaotic systems are exact: the equations describe them with full precision. What chaos means is that small uncertainties in our knowledge of the initial conditions grow exponentially over time, until they overwhelm any prediction. The structure is there. We just cannot see far enough through it.

Three bodies, gravity, and Newton's laws turned out to be enough to teach us that. The reason astronomy's simplest question has no clean answer is that the universe, even in its simplest cases, is more interesting than the mathematics we use to describe it.