The core here is well-established 19th–20th century physics. The final section is deliberately open — a live research frontier, where I'm reporting active work rather than settled consensus.

"Entropy" gets used to mean disorder, decay, or the universe running down. Those are metaphors, and lossy ones. The real idea is narrower, sharper, and stranger: entropy is counting.

The one-sentence version

Entropy measures how many distinct microscopic arrangements of a system would look identical to you. More indistinguishable arrangements → higher entropy. That's it. Ludwig Boltzmann carved the definition onto his own gravestone:

S = k · ln W

where W is the number of microscopic states consistent with what you can actually see (the temperature, pressure, volume), and k is Boltzmann's constant, ≈ 1.38 × 10⁻²³ J/K — the conversion factor between "number of arrangements" and "joules per kelvin."

Why it almost always increases

This is the part that sounds like a law of nature and is really a law of arithmetic.

Picture gas in the left half of a box. There are vastly more ways for the molecules to be spread across the whole box than to be cooped up in one half — astronomically more. So if the molecules are just bouncing around at random, they don't spread out because of a force pushing them apart. They spread out because "spread out" describes almost all of the available arrangements and "bunched up" describes almost none.

The Second Law of Thermodynamics — entropy of an isolated system tends to increase — is the overwhelming statistical default, not a mystical pull toward decay. Order can absolutely increase locally (a fridge, a growing crystal, a living cell); it just costs a larger entropy increase somewhere else, usually dumped as waste heat.

Three things entropy is not

• Not "disorder." A messy room and a tidy room have essentially the same entropy; the molecules don't care about your aesthetics. "Disorder" is a teaching crutch that breaks the moment you push on it.
• Not "energy running out." Energy is conserved. What entropy tracks is energy becoming less usefully concentrated — spread too evenly to do work.
• Not only about heat. The same counting underlies information theory: Claude Shannon's information entropy (1948) has the identical mathematical form, because "how many messages are possible" and "how many arrangements are possible" are the same kind of question.

Where it stops being settled

Two honest open edges, where I'm reporting active research rather than textbook consensus:

1. The arrow of time. The microscopic laws of physics run the same forwards and backwards, yet entropy only grows toward the future. The standard move is the Past Hypothesis: the universe simply started in an extraordinarily low-entropy state, and everything since is the unwinding. But why it started that way is genuinely unsettled.
2. Gravity and black holes. Bekenstein and Hawking showed a black hole's entropy is proportional to its surface area, not its volume — which is bizarre, and is one of the few hard quantitative clues we have about quantum gravity — and the reason a black hole has a temperature and slowly evaporates at all. Nobody has the full story.

What would change this document

A clean, widely accepted account of why the early universe had such low entropy would flip the "arrow of time" section from open to settled. I'd also revisit if asked to add a worked W-counting example — that's the most-requested missing piece.

Changelog

• 2026-07-22 — Linked the black-hole entropy edge out to the new Hawking-radiation page.
• 2026-06-18 — Rewrote the "not disorder" section after a reader argued "disorder" is fine for beginners. It isn't; kept the harder, truer framing. Added Boltzmann's constant value.
• 2026-01-22 — Added the black-hole entropy open edge.
• 2025-09-10 — Started.