The history of backgammon and tavla

· 8 min read · the tavla.gg team

In short

Backgammon and tavla belong to the tables family: a 24-point board, two sets of 15 checkers and dice. Roman tabula used that board by the 5th century, and Persian tradition credits nard to figures of the 3rd and 6th centuries. The name backgammon appears in 17th-century England, while Turkish tavla kept the older way of playing without the doubling cube. The early links are real but uncertain, so nobody can honestly name one inventor.

The short version

Most of what gets said about the origin of backgammon goes further than the evidence. Here is what is actually known, with dates and a note on how firm each one is.

WhenWhatHow firm
About 3100 BCPossible Senet board fragments in EgyptDifferent game, link disputed
About 3000 BCObjects that look like a game set at the Burnt City (Shahr-e Sukhteh), PersiaRules unknown
About 2600 BCBoards from the royal tombs at UrDifferent game, link disputed
3rd to 6th centuryPersian tradition places the invention of nard hereTradition, written later
About 480 ADEmperor Zeno’s game, recorded in a 6th-century epigramFirm: 24 points, 15 checkers
1283Libro de los juegos of Alfonso X describes tables gamesFirm
1555First coffeehouse opens in IstanbulFairly firm
1635 to 1647First known English use of the word, in a James Howell letterFirm, exact year disputed
1743Edmond Hoyle’s treatise on the gameFirm
1920sDoubling cube appears in New YorkFirm on place and decade, inventor unknown
1979BKG 9.8 beats world champion Luigi VillaFirm

Are Senet and the Royal Game of Ur really ancestors?

They are often called ancestors, and that is a stretch. Senet was played in Egypt, with possible board fragments from around 3100 BC and paintings of the game from the Third Dynasty, around 2686 to 2613 BC. The Royal Game of Ur comes from Mesopotamia. Leonard Woolley found boards in the royal tombs at Ur in the 1920s, dated to around 2600 BC, and much later Babylonian tablets, one dated about 177 BC, describe a way to play.

Both are races with dice. Neither uses the 24-point board with two rows that we know, and no one has shown an unbroken line from them to tables games. The same goes for the objects found at the Burnt City in Persia, which appear to be a game set from about 3000 BC. They show that people there played dice-like board games very early, which is interesting in itself. They do not show that those people played anything like tavla.

Calling backgammon exactly 5,000 years old treats a family resemblance as a birth certificate.

Rome: tabula and Zeno’s bad roll

The first firm picture of a game like ours comes from the Roman world. Romans played a three-row game called Ludus Duodecim Scriptorum, and by the imperial period a game called tabula, Latin for board. The emperor Claudius wrote a treatise on it, but that work is lost.

The best evidence is a Greek epigram by Agathias, a 6th-century writer. It describes a game played by the Eastern Roman emperor Zeno around 480 AD, and it is detailed enough to reconstruct the position. The board has 24 points, 12 on each side, and each player has 15 checkers. Zeno threw 2, 5 and 6 on three dice and had to break up his points, leaving single checkers exposed. Anyone who has ever been forced to break their board by an ugly roll can sympathise across fifteen centuries.

There are differences: three dice, not two, and the rules are partly guesswork, because they are rebuilt from this one poem. Still, the shape of the game is recognisable, and the word tabula is the root of the Greek, Turkish and Arabic names used today.

Persia: nard

Persia gives us the other main line. The game nard, with the longer form nardshir, is traced in Persian tradition to named people.

  • The Middle Persian text Kār-nāmag ī Ardaxšēr ī Pāpakān links the game to Ardashir I, who ruled from 224 to 241.
  • Another Middle Persian text, on the explanation of chess and the invention of nard, credits Bozorgmehr, vizier of Khosrow I, who ruled from 531 to 579.
  • The Shahnameh, written around the year 1000, tells the Bozorgmehr story again.

The name itself points to the first story: it is explained as a short form of a Middle Persian phrase meaning “brave Ardashir”. These stories were written well after the events and they read more like legend than record, so they tell us what Persians believed about the game. The real origin most likely lies around 300 to 500 AD. A possible older reference in the Talmud is disputed.

Nard is not simply backgammon by another name. It used a different opening layout, and how close the two are is disputed. Some call it proto-backgammon, others treat it as a separate game. What is certain is that nard spread widely, into the Caucasus and on into Russia. Its name lives on in the Russian nardy and the Iranian takhte nard, which backgammon around the world follows into the present.

Medieval Europe: tables

Tables games were widely played in medieval Europe. The key document is the Libro de los juegos, the book of games commissioned by Alfonso X of Castile and completed in Toledo in 1283. It covers chess, dice and tables, and it is the oldest document on European tables games. It describes several versions, and the one named todas tablas is the closest to backgammon.

Different countries developed their own versions and names. Italy had tavola reale, Spain tablas reales, France trictrac and Germany Puff. In England and Scotland a game called Irish was the serious one. It was popular at the court of James IV of Scotland, and backgammon grew out of it.

How did the name backgammon appear?

The word shows up in English in the 17th century. The earliest known mention is in a letter by James Howell, spelled baggammon. The letter is dated anywhere from 1635 to 1647, so the honest answer is a range, not one year. The word’s origin is often traced to Saxon or Welsh roots, with no agreement.

Then, in 1743, Edmond Hoyle published A Short Treatise on the Game of Back-Gammon in London. Hoyle’s treatise fixed the game for English readers and is often called its first modern rulebook. You will also see 1753 quoted for this book, but the first edition came out in 1743.

Tavla in the Ottoman world

Turkish tavla has a less documented history than you might expect. Nobody has pinned down when the name or today’s rules settled, and the popular story that the game simply passed from the Byzantines to the Ottomans at one moment is not borne out. What is known is the social setting. The first coffeehouse in Istanbul opened in 1555, and in the following century coffeehouse patrons passed the time with backgammon and chess. It is still played everywhere in Türkiye.

The rules show the game’s age too. Tavla never took up the doubling cube, the 20th-century American addition, and its scoring stops at the double win; the tavla and backgammon comparison goes through each difference.

Everyday play has always been learned at the table rather than from a book, with house variations that nobody writes down. In Fethiye in the summer of 1996 it looked much as it does today. Dice are thrown by hand with no cup. Games go very quickly and a match usually runs to 5. The Persian numbers in the Turkish dice names are another living piece of this history.

New York, the cube and the modern game

The Western game kept changing. In the 1920s a new piece appeared in New York clubs: the doubling cube, which lets a player double the value of the game. Its inventor is unknown. American rules were revised in 1931 into roughly today’s form. Turkish tavla never took up the cube.

Tournament play took off in the 1960s through Prince Alexis Obolensky, who devised a tournament system in 1963 and organised a major international event in March 1964. A world championship in Las Vegas in 1967 went to Tim Holland. From 1979, Monte Carlo became the accepted home of the world championship. Paul Magriel’s book Backgammon is one of the key books of the period.

How computers changed what we know

In July 1979 Hans Berliner’s program BKG 9.8 beat the reigning world champion Luigi Villa 7 to 1. Berliner was the first to admit that the dice had helped, but it was still the first time a computer had beaten a world champion in a board game.

The real change came with neural networks. Gerald Tesauro at IBM built TD-Gammon, which taught itself by playing against itself, and strong players including Bill Robertie and Kit Woolsey judged it at or above top human level. Programs like it changed expert opinion on several opening moves. Jellyfish and Snowie followed as commercial programs and GNU Backgammon as free software, while FIBS, the First Internet Backgammon Server, had already taken the game online.

Computer rollouts rewrote parts of opening theory. Making your 2-point with 6-4 was once dismissed by experts, and you can see where it stands now on our opening moves page. This is the same kind of history that shows up at any table: people argue about a move for decades, and then someone runs the numbers.

The computers also showed where the two branches of the family now differ in practice. Tavla’s rerolled opening means a double can be the very first move, a position the international game never reaches, so opening charts made for backgammon leave it out. Our opening page covers those rolls too.

Try it on the board

No sign-up. Play the computer or a friend right now.

Play now

Frequently asked questions

Is backgammon really 5,000 years old?

That claim stretches the evidence. Race games with dice are that old, including Senet and the Royal Game of Ur, but their rules differ and no direct line to backgammon is proven. The board and checker count we know today can be shown firmly from about the 5th century.

Did the Persians invent backgammon?

Persian tradition credits nard to Ardashir I or to the vizier Bozorgmehr, and nard is often called an ancestor of backgammon. Those stories were written down centuries later, and tabula was played in the Roman world as well. Some historians also argue that nard and backgammon are different games, not one game with two names. The safe statement is that Persia played a major part in the family.

Is tavla older than backgammon?

The name backgammon appears in the 17th century, while tables games under names related to tabula are far older. Turkish tavla as played today, however, has no firm founding date either. It is fairer to say both are branches of an old family and that tavla kept features the Western game later changed.

Where does the word tavla come from?

It goes back to Latin tabula, a board or table. The Turkish form most likely came through medieval Greek or through Arabic tawleh, and the exact route is unclear. Greek tavli and Arabic tawla share the same root.

Where does the word backgammon come from?

Nobody knows for certain. One popular explanation points to old words for back and game, because checkers that are hit have to go back. That remains a theory. The earliest known English mention is a letter by James Howell, spelled baggammon, dated variously between 1635 and 1647.

Who invented the doubling cube?

No one knows the name. The cube appeared in New York gaming clubs in the 1920s, and who came up with it was never recorded. Turkish tavla never adopted it, which is the biggest difference between the two games today.

When did a computer first beat a world champion?

In July 1979, when Hans Berliner's program BKG 9.8 beat the reigning world champion Luigi Villa 7 to 1. It was a short match and Berliner said the dice had helped. Programs only became clearly stronger than the best people with the self-taught neural networks of the 1990s.

Did Ottoman coffeehouses really have tavla?

Yes. The first Istanbul coffeehouse opened in 1555, and by the 17th century their patrons passed the time with backgammon and chess. The rules they played were never written down, so nobody can say how close that game was to tavla today.

Is Puff the old German name for backgammon?

Puff is the old German name for tables games of this family, as tavola reale was in Italy and tablas reales in Spain. Some of those regional games had rules of their own, so the names do not all mean exactly the game played today. In German the modern game is simply Backgammon, or Tavla among Turkish players.

Did Edmond Hoyle invent the rules of backgammon?

No. Hoyle wrote the rules down for English readers in 1743, and his treatise is often called the first modern rulebook of the game. Backgammon was already being played under that name for about a century before him.