Entropy

The universe's most fundamental tendency — and the reason anything happens at all.

Scroll to explore

What is Entropy?

Entropy counts how many different ways you could rearrange the hidden details of a system without changing what you actually observe. To understand it, you need two ideas:

Macrostate
The big picture you can measure from the outside — temperature, pressure, "how many heads." It’s what you see without looking at every detail.
Microstate
The exact arrangement of every particle or element that produces that macrostate — which specific coins are heads, which molecules are where.

Entropy measures how many microstates map to a given macrostate. More microstates = more ways to be disordered = higher entropy. Fewer microstates = rarer, more special = lower entropy.

Click to shuffle — watch disorder grow
Interactive

Macrostates & Microstates

Imagine flipping 4 coins. The macrostate is the summary: "how many heads?" The microstate is the exact sequence of every coin. Let's compare two macrostates:

Low Entropy

Macrostate: "All 4 Heads"

There is only one way to arrange 4 coins so they’re all heads. This macrostate is extremely specific — it has very few microstates.

H
H
H
H
HHHH
1 microstate → low entropy
High Entropy

Macrostate: "2 Heads, 2 Tails"

There are six different ways to pick which 2 of 4 coins land heads. Same summary, many arrangements — many microstates.

H
H
T
T
HHTT
H
T
H
T
HTHT
H
T
T
H
HTTH
T
H
H
T
THHT
T
H
T
H
THTH
T
T
H
H
TTHH
6 microstates → high entropy

This is why entropy increases. If you randomly shake or disturb a system, it's 6× more likely to land in the "2 heads, 2 tails" macrostate than "all heads" — simply because there are more ways to be mixed up than perfectly ordered. Scale this to 1023 molecules in a gas, and the high-entropy macrostates don't just win — they are overwhelmingly, astronomically more probable. Order isn't forbidden. It's just vanishingly unlikely.

Simulating 10,000 random 4-coin flips
Click to re-run simulation
S = kB ln W

Boltzmann's equation ties it together: S (entropy) equals kB (Boltzmann's constant) times the natural log of W (the number of microstates).

More microstates → higher entropy. It’s that simple. The "all heads" macrostate has W = 1, so S = 0 (minimum entropy). "2 heads, 2 tails" has W = 6, so S = kB × 1.79 — much higher.

Now imagine not 4 coins but 1023 gas molecules in a room. The number of microstates for an "evenly spread" macrostate is incomprehensibly larger than "all molecules in one corner." That’s why gas fills a room and never spontaneously empties one half.

Entropy Always Increases

Because there are overwhelmingly more ways for a system to be disordered than ordered, and because particles are constantly bouncing off each other in random directions, systems naturally drift toward higher-entropy states. Drop ink into water, and the molecules scatter in every direction — not because a law forces them to, but because almost every possible outcome is one where they spread out further.

There is no law that says the ink couldn't randomly bounce back together. It’s not forbidden. But the probability is so staggeringly small — like shuffling a deck of cards and getting them in perfect order — that you could watch for the entire age of the universe and never see it happen. The Second Law isn't really a "law" in the usual sense. It's a statistical certainty — the inevitable result of probability at enormous scales.

dS ≥ 0 (statistically guaranteed, not physically enforced)
10104
Current entropy of the
observable universe (in kB)
10120
Maximum possible entropy
(heat death)
13.8B
Years since the low-entropy
Big Bang
Ink diffusing — entropy increasing in real time
Click to drop ink

Order from Disorder

Here’s something surprising: the very process of entropy increasing can create pockets of order along the way. No life required — just physics.

Imagine sunlight hitting a pool of water. The sun heats the surface. Hot water is less dense, so it rises. Cooler water sinks to replace it. What emerges is a convection column — an organized, circulating flow that wasn’t there before.

This convection cell is locally ordered — molecules moving in a coordinated loop instead of randomly. But zoom out: the whole reason it exists is to move heat from the hot surface to the cooler depths faster. The structure accelerates entropy production. The universe “finds” that creating this organized pattern is the most efficient way to spread energy out.

The pattern is the shortcut. A still pool transfers heat slowly by random diffusion. A convection column transfers it much faster. The ordered structure is entropy’s tool for increasing itself more rapidly.

Sunlight → convection column → faster heat dispersal
Live simulation

Life: Entropy’s
Greatest Accelerator

Now look closer at that pool. The same convection currents that move heat also concentrate molecules. Nucleotides — the building blocks of RNA — get swept together by the flow, pushed into crevices and pores along mineral surfaces where they reach high enough concentrations to start linking up.

Convection concentrates nucleotides → RNA assembles
Live simulation

Convection concentrates

Thermal currents sweep dissolved nucleotides into pockets and mineral pores, creating locally high concentrations from a dilute solution.

RNA strands self-assemble

At high enough concentration, nucleotides spontaneously link into short RNA chains — a local entropy decrease, paid for by the heat released during bond formation.

RNA copies itself

Some RNA sequences can catalyze their own replication (ribozymes). Now the structure doesn’t just exist — it persists and multiplies, becoming an even faster entropy channel.

Life accelerates entropy

Each step — convection, concentration, assembly, replication — creates a more sophisticated engine for converting low-entropy sunlight into high-entropy waste heat. Life is the logical endpoint.

Same principle, scaled up. A convection column dissipates heat faster than still water. RNA dissipates energy faster than a convection column. A living cell faster still. Each layer of complexity is a better entropy accelerator — which is exactly why the universe “selects” for it.

Life is a function of an entropy gradient — whereby entropy increasing yields pockets of entropy-reducing local mechanisms that happen to self-replicate and evolve.

The Sun’s Gift

Earth receives ~1 high-energy visible photon from the Sun and re-radiates ~20 lower-energy infrared photons. Same energy, vastly more entropy.

🌱

Photosynthesis

Plants capture photons into tight chemical bonds (glucose) — a local entropy decrease paid for by increased universal entropy.

🔥

Metabolism

Your body breaks down ordered fuel molecules and radiates diffuse heat. You are a beautifully organized entropy-producing machine.

The Arrow of Time

Why does time seem to flow in one direction? The laws of physics are time-symmetric — a bouncing ball looks natural played forwards or backwards. But entropy gives time a direction. And the mechanism is more intimate than you might think: it lives inside your brain.

Entropy over time — the arrow only points one way

Memory Is
Entropy Increasing

Think about what happens when you form a memory. Your brain takes a disordered set of neurons and rearranges them into a specific pattern that encodes an experience. That sounds like a decrease in entropy — and locally, it is. But the process of writing that memory burns glucose, generates waste heat, and radiates infrared photons into the room.

The total entropy of the universe goes up every time you remember something. That means memories can only be created in the direction of increasing entropy — from a lower-entropy state to a higher one. You can’t form a memory of the future because the future is a higher-entropy state, and the thermodynamic cost has not yet been paid.

This is the deep insight: time doesn’t “move.” The laws of physics contain no arrow. What moves is you — your brain, recording one entropy-increasing snapshot after another, each one slightly more disordered than the last. You experience this chain of recordings as “the passage of time.” Time feels like a river carrying you forward, but really you are a series of states, each one only able to remember the lower-entropy states that came before it.

You don’t remember the past because it’s gone. You remember the past because forming that memory was an entropy-increasing process that could only happen in one direction. If entropy could decrease, you could form memories of the future just as easily.

Each memory recorded = entropy increased
Click to form a memory

Zooming Into
One Electron

Let’s zoom all the way in — past molecules, past atoms — to a single electron. In our everyday world, describing a thing is straightforward. A coin is either heads or tails. You can state which face is up and you’ve captured its state in one fact.

An electron should be similar. Tell me where it is, how fast it’s moving, and in what direction — and you know what it will do. Simple enough. But when we zoom in to actually look, something deeply strange happens.

The electron doesn’t have one position. It doesn’t have one speed or one direction. It has all of them at once. It’s as if you flipped a coin and instead of landing heads or tails, it landed on both sides simultaneously.

That’s hard to picture, because nothing in our world works this way. So instead of imagining a tiny ball, imagine a fuzzy cloud. The cloud doesn’t represent uncertainty about where the electron really is — the cloud is the electron. It is genuinely spread across many positions, many speeds, many directions, all at the same time. The cloud contains a bundle of possibilities, not a hidden definite answer.

The electron: a fuzzy cloud of possibilities
Click to “observe” the electron

When the Cloud
Becomes a Dot

Now imagine we point a tiny camera at the electron — we try to capture where it actually is. To do that, we have to bounce something off it — a photon, say. That photon comes back changed. It now carries an imprint of where the electron was. And that’s the moment everything shifts.

The interaction didn’t reveal a secret the cloud was hiding. The cloud genuinely had no definite position. But the photon forced a correlation — it entangled itself with the electron, and in doing so, stamped a definite outcome into existence. The fuzzy cloud of possibilities collapses into one definite electron with one position, one speed, one direction.

This isn’t some special property of cameras or conscious observers. Any interaction at all — a stray photon bouncing off, a neighboring atom feeling its electric field, a single molecule nudged by its presence — stamps a trace of the electron’s state onto the surrounding world. Each trace is a new correlation, and each correlation collapses the cloud a little further. The technical name is decoherence — the quantum fuzziness decoheres into classical definiteness.

Photon bounces off → imprint created → cloud collapses
Watch the interaction loop

And here’s the entropy connection — but why does spreading information mean more entropy?

Go back to the coin. It’s in superposition — both heads and tails at once. Now a marble rolls toward it and hits one face. If it strikes the heads side, the marble bounces left. If it strikes the tails side, it bounces right.

After the collision, you see the marble roll off to the right. You now know the coin landed tails — not because you looked at the coin, but because the marble’s direction recorded the outcome. The coin’s state got stamped onto the marble.

Before the collision, you only needed to describe one thing (the coin-cloud) to know everything. After it, the answer is spread across two things — the coin and the marble. You can’t fully describe either one without referencing the other. They’re correlated.

That’s more entropy. Remember: entropy counts microstates. When information is self-contained in one place, there’s one tidy description — few microstates, low entropy. When that same information is tangled across two objects, there are more ways to arrange the combined system that look the same from the outside — more microstates, higher entropy. Now scale it up: the marble hits an air molecule, which bumps another, which jiggles a photon — and the coin’s original answer fans out into billions of particles. Information disperses. Entropy skyrockets. Decoherence is entropy increase at the quantum level.

One interaction → information fans out → entropy rises
Click to trigger an interaction
S = −Tr(ρ ln ρ)

Von Neumann entropy: the quantum version of Boltzmann’s equation, where ρ is the density matrix describing the fuzzy cloud.

Before interaction: the electron’s cloud is a pure state. S = 0. All possibilities are coherent and self-contained.

After interaction: the electron is entangled with its environment. If you only look at the electron, you’ve lost information to the surroundings. S > 0. Entropy has increased.

Why Things Look “Normal”

Every object you see is made of trillions of particles constantly bumping into air molecules and photons. Their quantum clouds collapse trillions of times per second. That’s why your coffee cup doesn’t look fuzzy — decoherence happens so fast that everyday objects are always in definite states.

The Coin Analogy, Extended

A coin in superposition is both heads and tails. The moment anything interacts with it, a definite outcome gets stamped into the world. “Observing” isn’t magic — it’s just the first interaction that creates an irreversible trace. That trace is entropy.

Black Hole Entropy

Black holes have the highest entropy of any object their size. The Bekenstein-Hawking formula ties entropy to the event horizon area — hinting that information, entropy, and spacetime are deeply connected.

The Information Paradox

If decoherence is information spreading out, and black holes eventually evaporate via Hawking radiation, where does all that information go? This question sits at the frontier of physics.

But Wait — How Did
We Start So Ordered?

If entropy always increases, if the universe relentlessly moves from order to disorder, then the past must have been more ordered than the present. Trace that logic all the way back and you arrive at a startling conclusion: the Big Bang must have been a state of extraordinarily low entropy. But why? Of all the ways a universe could begin, an ultra-ordered start is overwhelmingly the least likely. So why did ours?

The universe: from an improbably ordered beginning to heat death

This is called the Past Hypothesis — the observation that the initial state of the universe had extremely low entropy. It’s not something we can derive from the laws of physics. It’s a boundary condition. The laws tell us entropy increases, but they don’t explain why it started so low. That question sits at the frontier of cosmology.

Smooth Beginning

Right after the Big Bang, matter was spread almost perfectly evenly — a hot, uniform plasma. That uniformity is low gravitational entropy. It’s the equivalent of all the gas in one half of a room, but for the entire universe. Gravity then pulled matter into clumps (stars, galaxies), increasing entropy as structure formed.

Inflation

One leading theory: a fraction of a second after the Big Bang, the universe underwent exponential expansion (inflation), stretching quantum fluctuations to cosmic scales and smoothing everything out. Inflation doesn’t fully explain why entropy was low — it pushes the question back to: why did the inflaton field start in that particular state?

The Multiverse Idea

Perhaps our universe is one of an enormous number, each with random initial conditions. Most start with high entropy and are boring — featureless heat baths. We find ourselves in one that started ordered because low entropy is a prerequisite for complexity, observers, and the question itself. Selection bias on a cosmic scale.

Cyclic Models

Maybe the Big Bang wasn’t the beginning. Some models propose the universe cycles through phases — expansion, contraction, bounce. Each bounce could reset entropy, or entropy could accumulate across cycles. The low-entropy start of our cycle might be the natural outcome of the previous one ending.

Everything in this entire journey — shuffling cards, ink diffusing, convection columns, RNA assembling, your brain forming memories, quantum clouds collapsing — all of it is only possible because the universe began in an improbably ordered state. We don’t yet know why. It might be the deepest unanswered question in all of physics.