David Parlett is a games inventor, historian and consultant and the author of many books on the subject. He specialises in the history of card games and has also published online article on the history of Poker, Gin Rummy and Blackjack/Pontoon for GameAccount.com.

Backgammon in the narrowest sense of the word denotes a game for two players, Black and White, who each start with their 15 pieces or table-men in the positions shown in the diagram above and move them respectively clockwise and anti-clockwise in accordance with the roll of two dice, the winner being the first to “bear” (carry off) all his men from his opponent’s starting-point.

The word itself means “back-game”, allegedly because the winner’s degree of victory depends on how far back his opponent is at end of play, or, less plausibly, “because the pieces, when taken, are put back” (Skeat’s Etymological Dictionary of the English Language). The word first appears in Cotgrave’s English-French Dictionary of 1611, and next in a letter dated 1645 under the spelling “baggamon”, which may represent the slovenly pronunciation characteristic of upper-class drawl.

However, the term Backgammon has a much wider range of application, in both space and time.

Geographically, it is also loosely applied to any game of similar type played in countries other than the English-speaking world. All are played on the 12-pointed board shown above, but otherwise exhibit a wide variety of different features. Some use three dice instead of two; some have different opening positions; some have both playing in the same direction rather than in opposite directions; some have different types of win; and so on.

Historically, it is also applied retrospectively to its immediate forebears, recorded in England as Tables from the 14th century to the end of the 17th, and before that under the Latin name Tabula from which Tables derives. Tables, however, is a generic term covering a variety of slightly differing games, including Irish, Doublet, Fayles, and several others. Its replacement by Backgammon in the 17th century most probably reflects the increasing popularity of the particular form of the game that we are most familiar with, which can be traced back to simpler precursor recorded in 13th-century Spain under the title Todas Tablas.

Even more expansively, it tends to be used for the whole family of similar games played in many of the oldest civilisations from the earliest known times. It really would be preferable to refer to that whole family as Tables, which is a much more generic name – appearing, for example, as Tavli in Greek, Tauli in Arabic, Tavole in Italian, Tafl in Scandinavian, Tawl(-bwrdd) in Welsh, and so on. Unfortunately, it would also be pedantic to insist on it.

(I once found myself in an embarrassing situation on this account. Having been introduced to a company as a games historian, and asked when and where Backgammon began, I replied “17th century England”, and was considered ignorant for not knowing it went back at least two thousand years. Whereas I interpreted the word Backgammon very specifically, for my interlocutor it meant any game of its type – in fact, he probably didn’t know there ever was any other game of its type.)

Therefore, when we speak of the history of Backgammon, we mean the whole history of Tables, and not just of the particular variety of it played in the English-speaking world. Even to speak of the family as Tables is to adopt a Eurocentric posture, as similar and probably ancestral games have long been widely and independently played throughout Asia.

The Backgammon Family

Given this range of application, it would be as well to draw attention to what all these games have in common by defining the family as a whole and comparing and contrasting it with other board game categories.

Backgammon, in its broadest sense of the word, is a race game. A race game is in which two or more players each move one or more personal pieces from Start to Home in accordance with the throw of random-number generators such as dice. The simplest race games are those in which each player moves only one piece, such as Snakes and Ladders (US: Chutes and Ladders). The fact that you have only one piece to move gives you no choice of play, therefore no possible strategy, and makes them games of pure chance.

In a more advanced group of race games, represented by the Indian games of Chaupar and Pachisi, and by the latter’s development into Ludo (UK) and Parcheesi (US), each player has four pieces to move and for the most part has freedom of choice as to how to apply the number obtained by casting. This introduces an element of skill and strategy, which in the Indian originals is increased by the fact that they are played by four in two partnerships of two, though each moves only his own distinctive pieces.

In the Backgammon or Tables family the number of pieces is increased to at least 15 per player (more, in some varieties), and the number of dice from one to two or three. This increases the skill factor to an extent where it outweighs that of chance, though not to as great an extent as that of diceless abstract board games such as Chess and Go. The centuries-old and virtually world-wide popularity of this type of game is therefore due to what is often described as its near perfect balance of chance and skill, which in turn makes it attractive to gamblers who like to be able to exercise control over the outcome of their play, and even enables the best to make a living from it.

Evidence has recently been unearthed to support the long-held suspicion that games of this type go back several thousand years. Certainly an ancestor of Backgammon, or at least a forerunner of it, is known to have been played in the Roman portion of classical antiquity. However, whether modern Backgammon represents an unbroken line of succession from it, or a development of a Persian game called Nard, is a question that remains to be resolved.

A helpful ground plan for this historical survey is that suggested by Willard Fiske in his uncompleted book misleadingly entitled Chess in Iceland (Florence, 1905). Fiske distinguishes three periods of development, to which, in my Oxford History of Board Games (Oxford, 1999) I attached an anterior fourth covering the classical and prehistoric varieties, producing the following scheme:

• Race games of classical and pre-classical antiquity, culminating in Tabula.
• Nard, from its invention or earliest appearance in south-west Asia prior to the 9th century AD.
• Tables, the European game in all its variations from around the turn of the first millennium.
• Backgammon, Tric-trac and other sophisticated games from the 15th century on.

Classical and pre-classical race games

The Royal Game of Ur. The oldest complete set of board-gaming equipment ever found involves a board consisting of 20 squares, some decorated with rosettes, and arranged in three horizontal rows of which the central row contains eight squares and the others four at one end and two at the other. Each player moved seven pieces in accordance with the throw of three tetrahedral (pyramidal) dice. It was unearthed from the royal tombs of Ur (about 105 miles north west of Basra) by Sir Leonard Woolley in 1926-7.

The Game of Twenty. A similar board to that of Ur, known as the Game of Twenty from the number of squares, is often found on one surface of Egyptian gaming boxes, with the better-known game of Senet on the other. Such a game is known from about the middle of the second millennium BC in Babylon, with later examples from Mesopotamia and Iran. Each player had five pieces, entered them from the single end of the arm, and aimed to get them on to the four squares on either side of the opposite end, thence to bear them off.

Senet was played throughout the period of Egyptian dynastic history, from the 30th to the 4th centuries BC. The board consists of 30 squares arranged in three rows of ten and the number of pieces per player varied from five and seven, eventually settling down to five in the New Kingdom (second millennium BC).

Shields, palms, dogs and jackals. There are many remains of a perhaps much older game played by moving pegs around 58 holes on a bilaterally symmetrical board, like a Cribbage board but artistically designed in various shapes such as a shield, a double axe-head, a palm-tree, and so on. It is variously known, from the shape of the board or its pieces, as the Shield Game, the Palm-Tree Game, or the Game of Dogs (or Hounds) and Jackals.

Grammai is the likeliest candidate for a Greek precursor of Tables. The games of ancient Greece are poorly known, but Murray (in A History of Board Games Other than Chess, Oxford, 1953) reasonably concludes ‘that the board games of Greece and Rome are affiliated to the older games of Egypt, Ur and Palestine, and that they reached Greece by way of the Mediterranean islands’. The game referred to as Grammai (meaning lines scored on a surface), later as Diagrammismos, though often mentioned in Greek literature, has left no adequate description apart from an inconclusive line drawing engraved on the back of a metal hand-mirror. All it shows is a board on which are inscribed twelve parallel lines at right angles to, but not touching, the longer sides of the board. Neither dice nor playing-pieces appear, though one of the players has one hand raised as though on the point of making a cast. It would certainly be possible to play some kind of Backgammon on such a board, but that’s the most that can be said for it. Almost certainly, however, it relates to the following entry.

Duodecim Scripta (“Twelve lines”), the Roman equivalent of Backgammon, was played throughout the Roman Empire up to and beyond the beginning of the Christian era and has left many artefactual traces, mostly of cubical dice and of boards scratched on hard surfaces, though hardly any of playing pieces. It is known that each side had 15 of these, as in modern Backgammon, and that three dice were thrown, as in many other varieties of Tables. The most distinctive difference, however, lay in the fact that it was played on three rows of 12 spaces. In this respect it differs from all modern members of the family, but appears to be the logical culmination of a whole series of ancient games played on three rows of various lengths, going right back to the Egyptian Game of Thirty Squares (perhaps, in some variations, expanded to thirty-six?), and, in a sense, to the Royal Game of Ur.

Tabula. At some time around the start of the Christian era the Romans came up with, or encountered, an apparent modification of Duodecim Scripta made by reducing the number of rows from three to two, still of twelve spaces each, and so producing a game substantially similar to Backgammon, if not virtually identical with it. This new game became popular in the 1st century AD, being especially favoured by the Emperor Claudius (r.41-54). “So addicted was his devotion to dice,” says Suetonius (in Robert Graves’s translation of The Lives of the Twelve Caesars), “that he published a book on the subject, and used to play, while out driving, on a special board fitted to his carriage which prevented the game from upsetting.”

Tabula is also the game which was primarily responsible for the gambling mania which swept Rome prior to its being declared illegal under the Republic. The fine for gambling at any other time except the Saturnalia was four times the stakes, although this law was only weakly and sporadically enforced. (From gammoned.com. It would be helpful to have a source cited for this statement.)

The word used by Suetonius for the game is Alea, which simply means dice (strictly speaking “die”, singular). This remained the word for it in early medieval Latin, but it everyday speech it soon became known as Tabula, which basically means a flat surface and, by extension, a table or gaming-board. As such, it passed into Byzantium and thence entered Greek as Table (two syllables), whence the modern Greek Tavli and other European variations on the word. Table is the subject of four epigrams by Agathias of Myrine, one of which concerns an embarrassing position fallen into by the Byzantine Emperor Zeno (r.475-481). The point of the embarrassment is considerably reduced, however, if we assume that the game is played in contrary motion (with players moving in opposite directions, as in modern Backgammon) rather than in parallel motion (following the same direction from start to home), which seems to be the case here. And which leaves us wondering whether Duodecim Scripta itself was a parallel or a contrary game.

This brings us to a period which as far as Europe is concerned is understandably designated the Dark Ages. It’s a period in which a tremendous amount of historical stuff was going on by way of population movements and language change, but of which tantalisingly little was recorded in writing, making it difficult to follow many strands of cultural history and especially that of games. St Isidore of Seville (c.560—636) mentions some such game under its older name Alea in his encyclopaedic work “Origines”, suggesting that it had not died out completely. But by this time the dark-age European back-gaming tradition was about to be regenerated by that of the more enlightened civilisation of Islam.

The Persian Nard

The Islamic contributor to the European games of Tables is or was the Persian game of Nard. The modern Iranian game still practised under that name is virtually identical with Anglo-American or international Backgammon, though it would be straining credulity to believe that the original was played in exactly the same way. All games evolve, and a significant mechanism of game evolution is the continual interplay of gaming ideas (“ludemes”) by means of borrowings and back-borrowings from one culture to another.

The name of the game first occurs in the Babylonian “Talmud” (300-500 AD) in its extended form Nardshir. Nard means a cylindrical block of wood and shir means lion, suggesting an origin in the design of the pieces used for playing it. It first appears in its shortened form in the Middle Persian romance “Chatrang-Namak” (“The invention of Chess”), a 7th century Persian story purporting to explain how Chess reached Persia from India.

Dewasaram, a great (but fictional, or at least unidentified) Indian raj, sends his vizier Takhtaritus to Nushirwan, the so-called Shahanshah (= king of kings) of Persia (531-578), together with many riches including an ornate Chess set with emerald and ruby pieces. These are all for Nushirwan, provided that he can explain the nature and use of the Chess set. If not, Nushirwan hardly merits the greatness he claims for himself, should return the riches with interest, and pay tribute into the bargain. Chess being then unknown to the Persians, Nushirwan is, to say the least, embarrassed. Fortunately, the problem is solved by the sage Wajurgmihr, who not only realises that this chatrang thing is a war-game, but also deduces the aim of the game and the movement of the pieces, and demonstrates his skill by beating Takhtaritus twelve times in a row. As if this were not enough, he devises a completely new game — Nard, the equivalent of Backgammon — and on behalf of Nushirwan takes it to the court of Dewasarm with the same challenge. Needless to say, no one in the Indian kingdom deduce its rules. Accordingly, “Wajurgmihr received from Dewasarm twice the tribute and revenue, and he returned in good health and with great ceremony to Iran.

If this story (repeated with variations in an 11th-century Arabic “Book of Kings”) establishes at least literary credentials for the Persian origin of Backgammon, it is perhaps reinforced by some recently discovered archaeological remains from Jiroft in south-eastern Iran. These include a low table whose surface is a gaming board consisting of three parallel rows of twelve circular playing cells formed by the coils of four snakes. An empty space in the centre, dividing the track into two wings, gives the whole thing the appearance of Duodecim Scripta, the three-row Roman ancestor of Alea/Tabula. If its tentative dating to the mid third millennium BC is accepted, it looks very much as if the ancestor of Backgammon goes back some 5000 years.

By their own traditions and all available evidence the Arabs adopted both Nard and Chess from the Persians following their 7th-century conquest of what is now Iran. That Nard came first is suggested by the fact that Islamic legal traditions allude to its existence in the Prophet’s lifetime, but make no mention of Chess before his death. A mention of Nard (there is none of Chess) in the Gamara, or Babylonian Talmud, is thought to have been inserted in 6th-century Syria, and an early reference to both occurs in an 8th-century letter from the Caliph Mahdi, rebuking the people of Mecca for such loose habits as the practice of playing these games. Despite such injunctions, they both came to play a prominent part in Arabic culture in the ensuing centuries. El’Adli, the first Arabic author of a practical work on chess in the mid 9th century, is also reported to have composed a book on Nard (now lost), and in the 12th century one Muhammad Sukaiker, of Damascus, wrote a work designed to prove the superiority of Chess to Nard.

The game also spread northwards in Georgia, to the Kalmuks of Central Asia, the steppes north of Astrakhan and into the Deccan, to China and thence Korea and Japan, as well as westwards, entering Europe through the Moorish conquest of Spain.

Oriental detour

Claims for India as the birthplace of Nard/Backgammon are not generally accepted, but several varieties of the game are described under the name Pasakakrida in the “Manasollasa”, and early 12th-century Sanskrit encyclopedia by a prince of the Calukya dynasty. A point of particular interest is that six different opening arrangements are detailed, one of which is that of what we now regard as the standard or traditional array. So far as I can ascertain, this is the earliest attestation of it. nevertheless, race games of this family have never been as popular in India as that country’s own native Chaupar.

The Chinese equivalent of Backgammon, Shuang-lu (“Double Sixes”), is known by name from the 6th century, when a History of the Southern Dynasty refers to a game played in 548 by the future Emperor Yuan with one of his ministers, and remained in China the classic of its type until the early 19th century. A 14th-century woodblock shows two players at the traditional wooden table, with raised lips around the edge, whose surface constitutes the playing board. The pieces – arranged at start of play in classic Backgammon formation – are not flat discs but tall and thin and vaguely bottle-shaped. Another distinctive feature of the Chinese game is that the 24 points are literally points (or, more elaborately, rosettes) rather than the elongated triangles of western Backgammon or the simple rectangles of ancient precursors, and that that the bar – the space separating the two rows of six on each side – is wider than in the west and is decorated with an elongated inverted crescent. The earliest Chinese board or table of this type, found at Astana in eastern Xinjiang, closely resembles one found in a 4th or 5th century tomb at Ballana in Sudan. The latter is assumed to be a Roman import and may represent a type widespread throughout the Roman empire. (Rosette-marked Duodecim Scripta boards are also known.)

Simpler boards of the rectangular-space type have also been unearthed, the oldest from a 6th-7th century Buddhist tomb in Xi’an. Other early Chinese games of this family include Wu-shuo (“Holding Spears”) and Chang-xing (“Long March”). These had died out by the 12th century, but many local varieties of Shuang-lu were still being played at this time. Some of them are described in a 12th-century treatise by Hong Zun (1120-1174), who gives this homely piece of observation:

“In the region of Yan (Hebei Province), teashops often provide boards, and each shop will have five, six, or more than ten. Players rent these boards with cash. This is like the inns and shops in Zongzhou (Henan Province) do with chess sets. Every Chinese or Khitan house-hold will have a backgammon board. Even when they travel by horse, they have attendants to carry the board. The dice are placed in a small leather pouch.” (From an article by Andrew Lo in “Asian Games; the Art of Contest”, N.Y. 2004.)

Hong also says “Shuanglu began in western India, and arrived in China during the Wei dynasty of the Cao family, thus reflecting a persistent (but not unchallenged) tradition as to its origin. Here it may be noted that the tall and conical tablemen are not typical of other Chinese games such as Wei-qi (= Go), but are so of India and Persia.

Shuang-lu also spread from China throughout the orient, becoming Ssang-ryouk in Korea and, in Japan, Sugoroku, or sometimes Ban-sugoroku to distinguish it from a later, unrelated game. Sugoroku is the oldest known board game of Japan and enjoyed great popularity throughout the medieval period until its eventual eclipse by western Backgammon from about the beginning of the 17th century. An 8th-century chronicle records Emperor Jito banned sugoroku “because, in the eye of the country’s highest administrative office, the game fostered gambling and those hooked on it would forget their works and filial duties”. It was much played by the aristocracy, and from the 12th century seems to have been administered by government officials. Besides being a gaming activity, Sugoroku served also an unusual ritualistic purpose in connection with childbirth. A 14th-century diary records a performance of the in a room adjacent to the delivery room of two successive royal consorts, in prayer for easy childbirth, and it is said that the ritual of rolling dice for a few days after childbirth was performed in an aristocratic household until early modern times. Towards the end of the 19th century, however, the Japanese government embarked on an active policy of westernisation, resulting (among so many other things) in the disparagement of Sugoroku as reminiscent of an outgrown feudalistic past. Western Backgammon has achieved particular prominence since the end of the Second World War.

Tables in Medieval Europe

Nard undoubtedly entered Europe through the 8th-century Moorish conquest of Spain, or through the Italian trading route, or both. Contact with the world of Islam was instrumental in ending the so-called Dark Ages of Europe. The post-millennial rise in cultural activity itself and the written records of it includes what now appears to be a renaissance in the proliferation of board games, to be followed in the 14th-century by the arrival of cards and card games, another import from the Arabic east. Games of the Backgammon type flourished throughout the European continent during this period, largely under the generic title of Tabulae in the written language (Latin) and its equivalents in the nascent national languages, such as Tables in French and English, Tavole in Italian, (Wurf-)Zabel in German, Tavli in Greek, and so on.

If we ask whether Tables represents the continuation of a European tradition going back to the Roman games of Tabula, Alea and Duodecim Scripta, or a direct offshoot of the Arabic version of Nard, the answer must surely be a combination of the two. There is no reason to believe that the Roman game died out completely during the Dark Ages, and it is more likely that with the introduction of Nard the new game – no doubt in several different varieties – simply assimilated itself into an existing tradition, influencing it to a greater or lesser extent at various points of contact. It is important to remember that Tables denotes, as it did at the time, not one fixed and unchanging game but a wide variety of similar games played on a board of 24 points with (usually) 15 white and 15 black tablemen and two or three dice – as indeed it still does, a fact nowadays obscured in the west by the cultural hegemony of Anglo-American Backgammon. Some of these games were played with three dice and all pieces either off the board or piled on to the ace-points, as in the Roman tradition; others with two dice and from the fixed opening position (like modern Backgammon) common to most eastern varieties of Nard and so presumably of the Persian game; and yet others with variations derived from various combinations of the two traditions.

Incidentally, we may note that the word “tabula”, originally denoting the board on which the game was played as well as the game itself, subsequently became “tabularium” (whence, for example, modern French tablier), indicating the board on which “tabulae” or “tablemen” were played. Thus the plural form Tables actually means tablemen, in much the same way that the French for Chess, “Echecs”, is plural because it means “the game played with Chess pieces”, and that American players refer to draughtsmen as checkers, original meaning “the pieces played with on a chequered board” – thereby morphing essentially square shapes into essentially round objects! (The name Merels (or Morris), denoting the other great board game of this period, also comes from a word meaning “gaming-pieces”.)

By a further – or perhaps parallel – process of development, table came to denote one quarter of a tabularium. In some context, Tables also implied the two quarters on each player’s side of the board (lengthwise), or the two quarters on the right or the left of the board, as separated by the (transverse) bar. The latter was particularly appropriate with the development of the typical European games compendium piece of furniture consisting of two hinged leaves with raised edges on one side. When shut into the form of a box containing chessmen, tablemen, and pieces for playing Draughts, Merels or Fox & Geese, the outer surfaces displayed a chequered board on one side and a Merels/Fox &Geese board on the other, these being games played “without (= outside) the tables”, as distinct from those games played “within the tables” – that is, Backgammon and its relatives – when the box was opened. The image produced by the hinged conjunction of two halves also gave rise to the practice of referring to the gaming board as “a pair of tables” in medieval English.

Our knowledge of medieval European Tables comes from a variety of sources, England being a particularly early one in this respect. In the Exeter Book, a 10th-century collection of Anglo-Saxon poetry, there is reference to two players “sitting at tables” (obviously the game, not the furniture), while a manuscript in the British Museum (Kings 13 A.xviii) describes eight games of the family as played around 1300 AD. The first of these, called Ludus Anglicorum (the game of the English), said to be the most popular variety and accordingly dealt with at greater length, is not Backgammon as we know it but that known as Emperador in Spain and Testa in Italy. It is played with three dice, or two and a nominal third throw of 6; each player starts with 15 men on the opposite ace-point; and there are positional grades of win known as “limpolding” and “lurching”. The basic set-up of this three-dice game (sometimes called the Long Game, Long Backgammon, Longammon, and suchlike) is more reminiscent of the Roman than the Arabic tradition.

More spectacular than this relatively obscure manuscript is the magnificent book of games commissioned by King Alfonso “the Wise” of Spain, which is generally dated to 1283. The beautifully illustrated Alfonso manuscript – hereinafter “Alfonso” for short – contains detailed descriptions of many varieties of games including Chess, Dice, Tables and Merels, some of them widely known, others evidently invented or designed specifically for the Spanish royal court. Some fifteen games of Tables are described, including Emperador (the Long Game) and something more akin to modern Backgammon under the title Todas Tablas, meaning All the Tables. (For reasons not satisfactorily explained, but possibly because the initial arrangement of pieces involves placing some of one’s own men in every quarter of the board, by contrast with games of the Roman tradition, in which they were entered from hand or piled on the ace-points.)

Other significant documents from France, Italy, and many other European countries testify to the widespread popularity of various games of Tables up to 15th century, when they found themselves rivalled on one hand by the introduction and proliferation of card games, which sidetracked those Tables-players who mostly enjoyed the gambling element, and on the other by substantial improvements to Chess and Draughts, which more effectively enticed those who preferred the intellectual challenge. These circumstances, perhaps along with a decline in population following the Black Death and its renewal with a substantially younger generation, resulted in a diminution of the number of varieties of Tables in favour of one or two major games associated with their own particular homelands. So, for example, the French developed their national game of Tric-trac, a game won by scoring points for achieving certain configurations, and the Germans for Puffspiel, a game of parallel rather than contrary movement.

From Irish to Backgammon

In England, the most popular 16th-century game of its type in the was the curiously named “Irish”, a title first recorded in 1588 and still in need of satisfactory explanation. It was to evolve into Backgammon from about the end of the century, the first recorded mention of that name occurring, as we have seen, in 1611.

Two major innovations distinguished Backgammon from Irish: the playing of doubles twice, and the introduction of degrees of win based on how far back one’s opponent was at the time. Specifically, you won a single game by bearing off all your men if your opponent had started bearing off, a double if you finished on a doublet throw, a triple or “backgammon”, if you finished before your opponent had borne any, and a quadruple if you did this on a doublet throw. Incidentally, while it seems plausible that the game took its name from the newly invented triple win made when the loser is “backward in the game”, or “backgammoned”, there can be no doubt that the meaning of the word is literally “back game”. The sometimes encountered suggestion that it derives from Welsh roots including “bach” (= little) is as ludicrous as the derivation of “penguin” from Welsh “pen gwyn” (= white head). “Gammon” may recall a side of bacon, but is more reliably interpreted as a typical piece of 17th-century happy-go-lucky spelling, in this case of “gamen”, the older form of “game”. By a later development, the increased win for a final doublet was abolished; the original backgammon (triple) was reduced to a double and renamed a simple gammon; and the backgammon triple win redefined into its present form.

The idea of degrees of win was not in itself entirely new. Earlier varieties of Tables included a double win or “lurch” made before the opponent had borne off any men. As this has been interpreted by some as an error arising from confusion with Cribbage, perhaps we should remember that it has already been encountered in the reference (above) to the Long Game, Ludus Anglicorum. (It ultimately derives, via French, from a German word meaning left-handed, or perhaps more appropriately “wrong-footed”.) Backgammon vied to some extent with its immediate forerunner, as suggested by this reference to it by a letter-writer of 1645 “Though you have learnt to play at Baggammon, you must not forget Irish, which is a more solid and serious game”.

Several games “within the tables” still played in the 17th century are described in three contemporary sources, namely Charles Cotton’s “Compleat Gamester” of 1674, Francis Willughby’s “Book of Plaies (= Games)” of about the same date, consisting of hand-written entries in a large notebook compiled over several years, and Randle Holme’s “Academy of Armory” of 1688. In addition, Thomas Hyde wrote (in Latin) a history of board games with particular attention to “Nerdiludus” – the game of Nard presumably for nerds – but concentrates more on the language and literature of the subject than on the actual rules of play.

Cotton describes, in an order implying relative degrees of status, the games of Irish, Backgammon, Tick-tack, Doublets, Sice-Ace, and Catch-Dolt, the last three of which he regards as “foolish pastimes... ridiculous to treat of”. Holme follows exactly the same pattern, and can be ignored, as he adds little of any value. Willughby – more expansive, more informative, and more enjoyable, since he was writing purely for his own amusement – starts with a description of the gaming equipment then goes on to describe Doublets, Tick-tack, Irish, and Backgammon.

Of Irish, Cotton says it is “an ingenious game and requires a great deal of skill to play it well, especially the After-game. It is not to be learn’d otherwise than by observation and practice.” Willughby goes into more technical detail, but, more amusingly, queries whether it is so called because the Irish invented this game, and comments “It is a proverb, An Irish man is never dead till his head be cut off”, possibly implying that the game is never safely won till you have borne your last man home.

Tick-tack is equivalent to the French Tric-Trac, and has its own peculiar terminology. As Cotton says, “Some play this game with Toots, Boveries, and Flyers; Toots is, when you fill up your table at home and then there is required small throws ... Boveries is when you have a man in the eleventh point of your own Tables and another in the same point of your Adversaries directly answering. Flyers is, when you bring a man round the Tables before your Adversary hath got over his first Table ...”). Willughby uses similar terminology, with Rovers in place of Flyers, and gives a list of “ridiculous names and phrases” of a rhyming nature used to designate particular throws of dice, rather like the language of Bingo-calling in Britain today. For example, “Trey-Ace “ is jocularly converted into “Trash and trumperie”, Cinque-ace into “Sing Alse and I’le Whistle”, and so on.

(Incidentally, the numeral terms ace, deuce, trey, quater, cinque, sice, used of dice throws and to a limited extent of playing-cards, are remnants of Anglo-Norman, the French of William the Conqueror and his lordly descendants for some 200 years following the illegal occupation of England in 1066.)

Doublets is more like Shut the Box than anything else, in that all the players do is enter their men in accordance with their casts, and then, without moving them, when all are in place, use the next series of casts to take them all off again. Willughby says “This is the most childish game at Tables in which there is nothing but chance and scarce any skill”.

Sice-Ace (Cotton) is said by Murray, in his “History of Board Games Other than Chess”, p. 123, to be an ordinary non-Tables dice game, yet Cotton’s description certainly makes it read like one.

A feature of Catch-Dolt (Cotton) is that it is played by four, and when one pair of partners have borne all their men they continue play, using subsequent throws to bear those of their opponents, and slapping their adversaries’ hands for each one they get home.

Neither of them gives pride of place to Backgammon. Willughby treats it last, and concludes this section by commenting “Of all games at Tables, Irish is counted the best”. But he could not read the writing on the wall. In 1694 Thomas Hyde describes Backgammon as the most widely played (“usitatissimum”) form of Tables, and he was undoubtedly right. By this time the term “Tables” had ceased to be current, and Backgammon had won the day.

Vicarious interlude

By the start of the 18th century, then, Backgammon had established itself as the first and foremost of English games played “within” the tables. But first and foremost with whom? At first, it continued to be with the gentry and “polite society”, as evidenced by the fact that Edmond Hoyle, following the success of his extraordinarily influential Treatise on Whist in 1742, next turned his attention to Backgammon. It has been said that he was the first to do codify its official rules of play, but in fact, as in all his works, he took it for granted that every one knew the rules and concentrated his attention on how to play it well, or at least successfully. Later, however, it seems to have gone into some sort of social decline, especially in town. This was the age when coffee houses were being replaced as centres of play by gentlemen’s clubs, where the principal gambling games were several varieties of cards together with EO, the ancestor of roulette. Players of more intellectual games were concentrating their attentions more on Chess and, at this time, Draughts.

The literature of the time tends to associate Backgammon with a sort of cosy gentility associated particularly with the rural clergy. Dean Swift, writing to a friend of his in the country, sarcastically asks “In what esteem are you with the vicar of the parish; can you play with him at Backgammon?” It is the preferred pastime of Oliver Goldsmith’s “Vicar of Wakefield” (1766), who gives us the following insight into his family amusements

“Walking out, drinking tea, country dances, and forfeits shortened the rest of the day, without the assistance of cards, as I hated all manner of gaming, except backgammon, at which my old friend and I sometimes took a two-penny hit. Nor can I here pass over an ominous circumstance that happened the last time we played together; I only wanted to fling a quatre, and yet I threw deuce ace five times running”.

And in 1735 Soame Jenyns pithily sums it all up in his delightful couplet

Choice books, safe horses, wholesome liquor / Billiards, backgammon, and the vicar.

In the opening years of the 19th century Joseph Strutt notes, in “The Sports and Pastimes of the People of England” –

“The tables, indeed, are frequently enough to be met with in the country mansions; but upon examination you will generally find the men deficient, the dice lost, or some other cause to render them useless. Backgammon is certainly a diversion by no means fitted for company, which cards are made to accommodate in a more extensive manner; and therefore it is no wonder they have gained the ascendancy.”

From literary sources, the reduced status of Backgammon can be perceived by the fact that Jane Austen, Strutt’s contemporary and herself the daughter of a country clergyman, delights in expatiating upon card games in all her novels, but barely mentions Backgammon.

20th century renaissance

The 20th century renaissance of Backgammon dates from the introduction of the doubling die in one of the American clubs in the 1920s. This stake-raising device has six sides numbered 2-4-8-16-32-64. Play starts for a single stake. Either player, when about to throw, may offer to double the stake by passing the die with its ‘2’ uppermost to the opponent’s side of the board. If the opponent resigns, the doubler wins a single game; if not, play continues at doubled stake. Subsequently, the doubled player, on his turn to throw, may offer to quadruple by passing the die with its ‘4’ uppermost across the table... and so on, following the same principle. It is still possible to win a “gammon”, thereby doubling the amount shown on the doubling die, but the thrice-valued “backgammon”, which gave the game its name when first introduced in the 17th century, is now generally disregarded, except by prior agreement.

Everyone speaks of this device as if it were a brilliant modern invention that would never have occurred to the ancients, but from the historian’s point of view it was nothing new. In the late 17th-century Francis Willughby described a process of doubling and redoubling the stakes at Ticktack, producing a variant he calls Vie Ticktack. And in the 12th-century “Manasollasa” advice is given, in the section on Pasakakrida, on positions within the game at which it is advisable to double.

The 20th-century invention was not, therefore, that of doubling in itself but rather that of doubling my means of an additional piece of equipment, the doubling die, which has the further merit, or at least discipline, of imposing an upper limit on the amount of doubling that can be entered into. At any rate, Anglo-American Backgammon became something of a fad on both sides of the Atlantic towards the end of the 1960s and really blossomed in the seventies. An annual world championship tournament was inaugurated in 1967 and has been held at Monte Carlo every year since 1976. In 1972 it was won by Oswald Jacoby, who, in collaboration with John Crawford, had two years earlier produced the first serious treatise on the game written in English since that of the original Hoyle in 1743.

(Serious it may have been, in the sense of historical and analytical, but by no means sombre in readability, enlivened as it is by such anecdotes as the following: “In all gambling games there is a loser’s syndrome which causes people to keep on when behind and to go further and further into the hole. The doubling feature in backgammon makes this syndrome far more dangerous. Psychologically, it is much easier to say “Let’s start the cube at 2 or 4” than it is to say “Let’s double the stakes.” In addition, this syndrome causes the loser to double too soon and to accept doubles which he should refuse. Almost anything can happen when a player gets caught up in a loser’s syndrome. The following incident occurred at one of New York’s leading clubs. A young member of modest means started to play for five dollars a point against a rather overbearing, wealthy older member. Luck favoured the younger man, but his opponent doubled and redoubled desperately, so that after an hour or so the youngster was five hundred dollars ahead. The loser insisted on starting all games at 2 and practically bludgeoned the young man into agreeing. The next step was to move the cube to 4, then to 8, and eventually up to 64. None of this did the loser any good. He just lost more, and the eventual tab came to around a hundred thousand dollars. The rich man, who had blustered and forced the stakes up, even against the young man’s steady protests, now said, “I don’t see why I should pay you. If you had lost, you couldn’t have paid me.” Then he offered a five-thousand-dollar settlement. The opponent, who had quietly stood the man’s insults and had behaved most properly throughout, now replied, “I know I couldn’t have paid a hundred thousand, but I could have paid everything I might have lost. You kept doubling to get even, and now you must pay the full amount you lost.” We are pleased to report that when the rich man had time to cool off, he did pay in full. Not that we approve of such high-stake gambling, but we do believe in paying gambling debts.”)

The 1970s also saw the beginnings of computer analysis of the game and the development of game-playing software. Indeed, Backgammon has been one of the greatest success stories in this branch of the electronic world. It lends itself to analysis more readily than Chess in several respects, for example, that it is linear rather than two-dimensional, that a completed game can never end in a draw, and that the analysis is essentially arithmetical. The earliest serious attempt at a backgammon program was that of a group led by Hans Berliner group culminating in the BKG position evaluator. In 1980 it beat world champion Luigi Villa, though it has been suggested that its success was due more to the shortness of the match and Villa’s apparent unwillingness to enter into it as seriously as he should have done. In the 1990s Gerald Tesauro created TD-gammon, which narrowly lost a match against former world champion Bill Robertie by 40-39. Tesauro’s estimated that TD-gammon was significantly stronger than BKG by about 0.35 points per game and approximated current human world-champion level. Its strength was due to its ability to With an input of the board position combined with basic backgammon knowledge and a random initialised network, TD-gammon trained itself on 1.5 million games of self play, resulting in world-class level of play.

The fruits of all this technology are to be seen (a) in a proliferation of readily available software programs enabling the home player to develop skills well in advance of those displayed by winners of the earliest world championships and (b) in the enormous popularity of Backgammon for casino, championship and private play online, in which respect it ranks with Poker and Blackjack.

The electronic renaissance of international (i.e. Anglo-American) Backgammon has by no means diminished the popularity of other games of the same family. Acey-Deucey retains a following in America; the French still play Trictrac and the Germans Puffspiel. And anyone who has ever visited south-eastern Europe and the middle east will have seen with their own eyes the continuing popularity of a multitude of national variants such as Tavli in Greece, Moultezim in Turkey, and, further north, of Narde in Russia. From one website devoted to Backgammon variants you can link to details of Ace-Deo, Ace-Mid Switch, Acey-Deucey, American Acey-Deucey, Backgammon to Lose, Backgammon 1931 Rules, Backgammon 1969 Rules, Backgammon 1970 Rules, Blast Off, Blocking Backgammon, Chasing the Girls, Chouette, Crazy Narde, Domino Backgammon, Doublets, Duplicate Backgammon, Dutch Backgammon, Eureika, European Acey-Deucey, Fayles, Fevga, French Backgammon, Gioul, Grande Trictrac, Grasshopper, Greek Acey-Deucey, Greek Backgammon, Gul Bara, Handicap Matches, Hyper-backgammon, Irish, Jacquet, LongGammon, Mexican Backgammon, Misere Backgammon, Moultezim, Nackgammon, Narde, Never-Finishing Game, Old English Backgammon, Pin Game, Plakoto, Plakoto Express, Poof, Portes, Propositions, Roman Backgammon, Roll-Over, Rosespring Backgammon, Russian Backgammon, Shesh Besh, Snake, Swedish Tables, Tables, Tabula, Takhteh, Tapa, Tavla, Tavli, Tawula, Tourne-case, Trictrac, Turkish Backgammon, Two Rolls versus Choice.

Nor does exhaust the list of possibilities as Backgammon enthusiasts exercise their imaginations by devising ever-more ingenious variations. Some of these will be found posted from time to time on the newsgroup.