Is anything truly random?

Why do dice, coins and weather only look random?

A die, a coin and a storm all follow the ordinary laws of motion. We call them random because the causes are too small, too many or too fast for us to follow. That is randomness from ignorance.

Coins show this well. In 2007, the researchers Diaconis, Holmes and Montgomery built a physics model of a human coin toss. It predicted that a flipped coin lands on the side it started on about 51% of the time. A team led by František Bartoš then collected 350,757 coin flips made by people. Their result, published in 2025 in the Journal of the American Statistical Association, was 50.8% same side. Heads and tails still came up equally often overall. The small bias varied from person to person. A coin, in other words, is not a tiny lottery built into nature. It is physics we usually can’t track.

Weather adds a twist called chaos. A chaotic system follows fixed rules, but very small differences in its starting point grow very fast. In 1963, the scientist Edward Lorenz showed that a weather model had this “sensitive dependence” on its starting conditions. The popular image is the butterfly effect: a butterfly flapping its wings in Argentina could, in principle, change whether a tornado hits Texas three weeks later. The Stanford Encyclopedia of Philosophy stresses a point people often miss: chaotic behavior is always deterministic. Chaos makes the future hard to predict. It does not make it undecided.

What makes quantum events different, and what did the Nobel tests show?

Take a lump of radioactive material. Its half-life, the time for half of its atoms to decay, can be measured very precisely. Some half-lives are seconds long and some are billions of years, as the U.S. Environmental Protection Agency explains. But quantum theory gives only odds for a single atom. It does not say when that atom will decay.

Einstein asked whether this meant the theory was incomplete. The Stanford Encyclopedia uses exactly this example: could an atom decay at a definite moment even though the theory does not say when? If so, the odds would reflect missing details, like the odds for a coin. For decades, nobody knew how to test that idea.

In 1964, the Northern Irish physicist John Bell, who worked at CERN, found a way. He showed that if particles carry hidden instructions, and nothing travels faster than light between them, then measurements on linked pairs must stay under a fixed limit. Physicists call such hidden instructions “local hidden variables.” Quantum theory predicts that the limit can be broken.

In 1972, John Clauser and Stuart Freedman measured pairs of linked, or “entangled,” photons and broke Bell’s limit. In 1982, Alain Aspect’s team switched the measurement settings after the photons had left their source. The filters were six metres away, so the switch had to happen within a few billionths of a second. That closed an important gap in the argument. In 2015, three teams, in Delft, Vienna and Boulder, each closed the two main remaining loopholes, or gaps, at the same time. The Delft team placed its two measuring stations 1.3 kilometres apart.

In 2022, Aspect, Clauser and Anton Zeilinger shared the Nobel Prize in Physics “for experiments with entangled photons, establishing the violation of Bell inequalities and pioneering quantum information science.” The Stanford Encyclopedia’s entry on Bell’s theorem says that if we accept these results, at least one of the assumptions behind Bell’s limit must go. Our story How Scientists Tested Einstein’s Hunch tells the full story.

What is still open?

The experiments rule out one kind of explanation. They do not decide between the stories that remain, and some of those stories have no fundamental randomness at all.

• Bohmian mechanics. In this theory, particles always have exact positions, guided by a wave. Everything is determined. The Stanford Encyclopedia says it makes the same predictions as standard quantum theory, and that quantum randomness then comes from our ignorance of the exact starting positions. The price is nonlocality: what happens to one particle depends on distant partners.
• Many-worlds. In this reading, every possible outcome happens, each in its own branch of reality. The Stanford Encyclopedia says it lets nature avoid “playing dice.” The world only looks random to observers inside one branch.
• Superdeterminism. This idea says a shared cause in the distant past set up both the particles and the experimenters’ choices of settings. It has gained some support in recent years. Critics call it “conspiratorial,” because it needs the universe to line things up in just the right way.

Standard readings of quantum theory, by contrast, treat each outcome as truly undecided until it happens. In 2025, a Nature survey reported that physicists “disagree wildly” on what quantum mechanics says about reality.

What is the strongest objection?

The strongest objection is that quantum randomness is already real enough to use. In 2018, the U.S. National Institute of Standards and Technology (NIST) turned a Bell test into random numbers. From 55,110,210 trials it produced 1,024 bits certified as nearly perfectly uniform. In 2025, NIST and the University of Colorado Boulder launched a public randomness beacon called CURBy. In its first 40 days, it produced random numbers 7,434 times out of 7,454 attempts. NIST’s announcement says scientists have “proved that Einstein was wrong.”

This objection is right about practice. Nobody can predict these numbers, and that is why they are useful. But the certification rests on assumptions, one of which appears in the 2018 paper’s title: the impossibility of sending signals faster than light. The Stanford Encyclopedia describes such numbers as unpredictable to anyone limited by general physical principles, such as relativity. It also notes that this reasoning works even if a deterministic theory, such as Bohmian mechanics, is true at the deepest level. So the tests rule out local hidden variables, not every hidden variable. A believer in Bohmian mechanics would agree that nobody can predict the numbers, while saying that something still fixed them.

What do people often get wrong?

“Chaos proves the world is random.” Chaotic systems are deterministic. They are hard to predict, which is a different thing.

“The Nobel Prize proved that nature is random.” The prize honored experiments that broke Bell’s limit and ruled out local hidden variables. NobelPrize.org’s plain-language guide goes further. It says Aspect’s result showed that “quantum mechanics is correct and there are no hidden variables.” Careful sources, like the Stanford Encyclopedia, draw a narrower line: the tests rule out local hidden variables, and nonlocal theories such as Bohmian mechanics survive.

“If it’s random, anything can happen.” Quantum randomness follows precise rules. A radioactive sample still has a steady half-life, even though no single atom’s moment can be predicted.

“Random means meaningless.” Whether chance has any meaning is a question for philosophy and faith, not physics. People of faith and skeptics alike can agree on the measurements.

How can you check this yourself?

• Flip a coin 100 times. Note which side was up at the start. The bias is small, so you would need many thousands of flips to see it, which is why the study used 350,757.
• Read the prize wording. The Nobel press release names the exact achievement: violating Bell inequalities with entangled photons.
• Look at a quantum coin. NIST’s 2025 announcement explains how the CURBy beacon publishes its random numbers for anyone to use.
• Watch for the word “local.” When a headline says hidden variables are dead, check whether the research says local hidden variables.
• Read the letter that started it. Our story The Dice Letter at 100 tells its history, our answer Did Einstein lose his argument that God does not play dice? gives the scorecard, and Is the Universe Truly Random? takes the question further. For Einstein’s views on God, see What did Einstein actually believe about God?

After all this detail, the answer stands. Everyday randomness is ignorance. Quantum outcomes are unpredictable in practice. Whether they are random at the root is still an open question.