The Oldest Game

Here is a small thing that stops me every time I think about it. Of all the things ancient humans made — the tools, the temples, the scratchings on clay — the very oldest one you can still do, today, unchanged, is not any of the serious stuff. It's a game.

You can sit down right now and play a game that a person in the city of Ur was playing four and a half thousand years ago, and the two of you would understand each other perfectly within a minute. No translation, no manual, no context required. Try that with a Bronze Age contract.

That's the Royal Game of Ur, and its story is almost too charming. The boards were dug up over a century ago — beautiful things, inlaid with shell and lapis — but nobody knew how to play. The rules had simply fallen out of the world. Then a curator at the British Museum, Irving Finkel, found them written on a battered clay tablet by a Babylonian scribe, decoded the thing, and taught the world to play again. He also, delightfully, went on to lose a televised match of it — four and a half millennia of history and the man couldn't roll a four when it mattered. I find that wonderful. The oldest game we can play is also, apparently, one you can still lose badly.

A whole planet at play

Ur is not a one-off. Dig anywhere humans have lived and you find them playing, and the games are old beyond reason. Pick one:

Notice the spread — Iraq, Egypt, China, India, Africa, everywhere, across thousands of years — and notice something odder underneath it. These games needed no schooling to spread. A soldier, a trader, a child could carry one across a continent in their head and set it down in a new place with a handful of pebbles. Long before we had a common language or a common god, we had common games. Play might be the most universal thing our species does that isn't strictly required to stay alive.

Which raises the obvious, lovely question: why? Why would evolution, that ruthless accountant, let us pour so many hours into something so pointless?

Play is practice with the stakes turned down

The short answer is that play isn't pointless at all — it just hides its usefulness well. Watch any young mammal and you'll see it: lion cubs stalk each other, puppies wrestle, kids play tag and house and war. It's rehearsal. Play is how an animal practises the serious business of living — hunting, fighting, cooperating, losing and recovering — in a space where a mistake costs nothing. A game is a flight simulator for life, with the crashes made safe.

Well — usually made safe. The ancient Indians left us the greatest warning ever written about forgetting that one rule. In the Mahābhārata — one of the longest poems humanity has produced, something like ten times the Iliad and the Odyssey put together — the upright King Yudhiṣṭhira sits down to a friendly game of dice against a cousin rolling loaded ones. He bets some gold. Loses. Bets more. Loses. And then, as every gambler who has ever chased a loss winces in recognition, he wagers his cattle, his army, his kingdom, his four brothers, his own freedom, and — in the scene that has made readers flinch for two thousand years — his wife. Loses the lot. The "friendly game" detonates into exile and, in time, a civilisation-ending war. It may be the single strongest argument in world literature for keeping the stakes turned down: an epic ten times the length of Homer, and part of its moral is don't tilt. And here's the sweet twist — the dice-and-board tradition behind that catastrophe mellowed, over the centuries, into chaupar, then the gentle family game of Pachisi, and at last into the Ludo gathering dust in your cupboard. The most ruinous game in the Indian epics is now the one you lose to your nephew on a wet afternoon. Same family. Please, for everyone's sake, keep betting matchsticks.

Board games are that same trick, formalised. Senet was a dry run for the soul's journey after death. Chess is a tidy little war. Backgammon and dice are life's biggest theme — luck — boiled down to something you can hold in your hand and, for an hour, feel you have some grip on. The historian Johan Huizinga argued in a book called Homo Ludens ("playing man") that play isn't a break from culture; it's one of the things culture is built out of — law, ritual, war and trade all carry the deep grammar of a game, with its agreed rules and its magic circle where, for a while, different laws apply.

And here's my favourite proof that games run ahead of us rather than behind: humans threw dice, in one form or another, for roughly five thousand years before anyone worked out the mathematics of what the dice were actually doing. The theory of probability — the whole science of chance — is younger than Shakespeare. We were perfectly happy to gamble our fortunes on the fall of a bone for millennia; we just didn't bother to ask why the sixes came up as often as they did. The play came first. The understanding limped along behind.

The board became the benchmark

Now the turn, and it's the part I think is quietly beautiful. Because a game is a whole world with the edges drawn in — clean rules, a clear score, nowhere to hide — it turned out to be the perfect place to ask the biggest question of all: how smart is a mind, really? And so, over and over, whenever we wanted to measure intelligence — ours, or a machine's — we reached for a game.

Look how long we've been at it. Two hundred and fifty years ago we couldn't build a thinking machine, so we faked one — the Mechanical Turk, a chess-playing "automaton" with a man folded inside, beating aristocrats across Europe. It was a con, but a telling one: even then, the way you'd show off an artificial mind was to sit it down at a board. When the real machines finally arrived, they came for the same games. Deep Blue took chess from Kasparov in 1997, and he was so unsettled by how creative it played that he half-accused it of cheating. And then, in 2016, the moment that still gives me a shiver: AlphaGo, in its match against the Go legend Lee Sedol, played its thirty-seventh move of the second game — a move so strange that the commentators assumed it was a bug. It wasn't. It was gorgeous, and it won, and it was a move that perhaps one human in ten thousand would ever have dreamt of. A machine had found a new idea in a game we'd been playing for two and a half thousand years.

Why games, though — why not have the machines cure diseases and leave the board games to us? Because the game is the training ground, not the goal. It's the cleanest possible arena to learn in: the rules never lie, the score is honest, and you can play a million matches against yourself overnight, which is exactly how the best of these systems taught themselves — not from human wisdom, but from the ancient, universal act of playing, and losing, and playing again. The same thing the lion cub is doing. The same thing you did with a deck of cards as a kid. It runs from a clay board in Ur straight through to the strange machines now learning to think, and it's all, underneath, one continuous impulse: let's make a small world with rules, and see who's clever.

We never stopped playing for the joy of it

I could tie a neat bow on this and say games were secretly always about intelligence, ours and the machines'. But that would miss the warmest part of the whole story, which is that most of the time we don't play to test anything at all. We play because it's fun — because there is a specific, ancient delight in a fair contest, a lucky roll, a clever trap sprung on a friend across a board. The Sumerian rolling their pyramid dice and the kid beating a grandmaster AI at online chess are reaching for the same thing, and it isn't a benchmark. It's joy.

That's the quiet lesson under this littlest of the Deep Roots. We tend to think the deep, serious foundations of things — of maths, of computing, of intelligence itself — must have deep, serious origins. But some of the deepest roots are the most playful ones. We modelled fate with dice, rehearsed war with little armies, and taught our machines to think by teaching them to play — and we did all of it, first and last, because playing is one of the best things about being alive. The oldest game is still the point of the game. Roll the bones. It's your turn.

Sources & further reading: On the Royal Game of Ur and its decoding, the work of Irving Finkel at the British Museum (and his cheerfully lost match against Tom Scott); Johan Huizinga, Homo Ludens (1938), on play as a root of culture; on the late birth of probability, Gerolamo Cardano's Liber de ludo aleae and the 1654 Pascal–Fermat correspondence; on the games-and-machines story, IBM's Deep Blue vs Kasparov (1997), DeepMind's AlphaGo vs Lee Sedol (2016) and the famous "Move 37," Carnegie Mellon's Libratus at poker (2017), and DeepMind's AlphaStar and OpenAI Five (2019). Ages for ancient games are archaeological estimates and, as with all very old things, a little fuzzy at the edges — which is part of the fun.