The game pieces moved on game boards in antiquity were called pebbles (yh=foi, Latin calculi). They were advanced, in most or all such games, in accordance with the fall of three (later two) dice (ku/boi: see kappa 2602), Latin tesserae), shaken from a small 'tower' fluted to prevent cheating (Martial 14.16; see delta 748, tau 7). These were, at least usually, gambling games, dependent not only on the luck of the dice but on the skill and experience of the players.
As a synonym for 'pebbles' (cf. Pollux 9.97) 'pessoi' (pessoi/, but sometimes pessa/, Sophocles, Euphorion fr.61 Powell) seems to refer to pieces shaped for stacking on points, lines or intersections of these boards, to such stacks (perhaps the lines or squares on which they stood: Eustathius 1396) and to games requiring stacking. This entry, together with pi 1391, is part of the evidence for this hypothesis. It is improbable that ku/boi in the definition here means 'dice' (but see Lamer 1938-9), or that it implies that playing pieces could be used as dice (cf. Suetonius, 'There is a clear distinction among the ancients between dice and pessoi'). The explanation comes from the architectural uses of the word pessoi (LSJ III). Strabo (16.1.5) describes the foundation blocks of a vaulted arch in the Hanging Gardens of Babylon as cube-shaped, hollow, earth-filled pettoi (= pessoi). Procopius (Buildings 1.1.37) uses pessoi for the four interior piers, made of shaped, fitted stone blocks, that support the two vaults of St. Sophia (cf. pi 1645). Whether the metaphor comes from architecture to board-playing or from board-playing to architecture, the inference of the parallel is reinforced by the Suda entry. Pessoi were stackable in shape, presumably cubic.
The Greeks and Romans played two different sorts of such board-games. The race-game has clear parallels to modern tavli, tric-trac and backgammon and is probably their ancestor (Oxford History of Board Games 70-73, with Egyptian parallels, as first suggested by Plato, Phaedrus 274C-D; cf. Suetonius pp. 64, 149-50 Taillardat). This game may have been first known as Grammai 'Lines' or Diagrammismos (so Austin, 1940, OHBG 70), or as Grammata (Plato loc.cit.). In Latin it was known as XII Scripta, later Alea. The Greeks later adopted for the game the Latin word tabula 'board' as ta/blh or Tabla (see tau 7), from which the modern Greek tavli is derived. The game was originally played with three dice on a board with 36 points on three lines, and the pessoi entered on the middle line. Their motion, circulating on the 12 points of this line, then on the 12 points in the reverse direction and finally on those of the final line in the original direction, was compared to the motion of the sun, moon and planets from East to West through the 12 signs of the Zodiac (tau 7), and boards may sometimes have been designed in the same fashion as the astrological chart of these motions (pli/nqia, see LSJ III 3; pesseuth/rion, see LSJ; cf. tau 7). Archaeologists have found many of the Roman boards and the 36 points are usually given letters of the alphabet, forming 6 relevant 6-letter words. About the time the names Tabla and Alea were adopted, the board was reduced to the modern 24-point game with two dice, probably the reason for the new names.
The poet Agathias describes a game in which the Emperor Zeno, apparently in an unbeatable position, made an unlucky throw of three dice on a 24-point board and lost his advantage (Greek Anthology 9.482). This situation is elucidated by Becq de Fouquieres and Austin (1934, cf. OHBG 72-3). The colorful list of dice throws given by Eubulus in his Dice-players (*Kubeutai/, fr.57) and Pollux (7.204-5; cf. Lamer 1945-58, and kappa 2602) includes moves in board-games resulting from the dice thrown; cf. the list for tric-trac moves in Acad. univ. Jeux (56ff.).
The war games (OHBG 234-7) seem to have developed and changed over the centuries and to have been played on boards of varying sizes (and hence number of lines) and with a varying number of pebbles or pessoi. Sophocles knew one as Pente Grammai and attributed its invention to Palamedes (Nauplius fr.429P = 393N). King Pyrrhus of Epirus developed a game of military tactics. The war game at Rome was known as Latrunculi, and a good player at this extraordinarily hard game of strategy, construction of 'camps', advance and retreat was praised in poetry (Calpurnius Siculus, Laus Pisonis 190-208) or on his epitaph (CIL 13.444). Several reconstructions have been attempted (see OHBG 236-7, citing one by Schaedler), but are unconvincing in the light of ancient evidence. Its play, and the role of dice in it, can be deduced to some degree from descriptions in Ovid, Tristia 2.475-80 (see S.G. Owen's edition and notes), the Laus Pisonis, fragments of Suetonius's lost On Games among the Greeks (edited with useful notes by J. Taillardat, 1967), Pollux 9.97-8, and a renaissance treatise and poem by Celio Calcagnini of Ferrara (p.296-300).
Martial has a pair of epigrams contrasting the play in XII Scripta, where a throw of double 6 allows one player to remove his opponent's 'blot', and Latrunculi, where one player's unprotected 'latro' is trapped and may be about to become his opponent's piece (14.17, 18). The center line was known as the 'sacred line' and the immovable '(sacred) stone' (li/qos, yh=fos) on it was, like the king in chess, especially to be protected. To play it was an admission of virtually certain defeat. This gave rise to the widespread proverb 'to move the stone from its line' for acting in desperation on the face of defeat (Alcaeus fr.351 Liberman = 82 Bergk = 306 Lobel-Page; Theocritus, Idylls 6.18 and scholia; scholia to Plato, Laws 820C; cf. Suetonius pp.65-6, 152-5 Taillardat). This game was still played (with moves determined by dice throws) in 16th century Ferrara. Calcagnini compares it favorably to XII Scripta and chess (which he calls Ludus Scaporum), as requiring more complex military tactics. It must have died out shortly after, perhaps with the rise of the courtly parlor games favored by Castiglione in The Courtier over difficult games such as chess (II xxxi).
Becq de Fouquieres, Austin (1940) and OHBG all identify the name of this game in Greece as Petteia and accept the distinction made by Hesychius (2010.22) between Petteia as a game where the moves are played without dice (cf. Plut. Moralia 427F, where the die is removed 'as in the game of pebbles') and Kubeia as the same game but with moves determined by the throw of the dice. This would make sense of Plato, distinguishing Kubeia, Petteia and 'Grammata' (?Grammai, the backgammon game) and of the passage in Philo Judaeus where 'fortune plays human affairs forwards and backwards' (see under pi 1391); for pieces in this game move in either direction but in Tabla games, like the heavenly bodies, always East to West. Plato describes it as requiring great skill (Republic 333B, 374D, 487B; Polybius 1.84; etc.; cf. pi 1391). In addition poleis, 'cities', were a feature of this game (cf. allusions in Plato and in Aristotle, Politics 1253a), in the sense of organized formations blocking the advance of the opponent, and may also have been a name for one version of the game (scholia to Plato, Republic 422C; Austin, 1940, pp. 263-6; Taillardat 154-5). But the entry at pi 1392 and other evidence suggest that Petteia was also used in general for other Greek board games using pessoi. Indeed poleis as formations are well attested in both backgammon and war games.
No doubt attempts will continue to reconstruct the lost game of Latrunculi and the 36-point game of backgammon, which were clearly as addictive to the ancients as tric-trac was in the courts of Louis XIV and his successors or tavli is in the eastern Mediterranean today. The Romans had laws against such gambling, but from Augustus onwards the emperors were inveterate gamblers with dice and board games. Toner analyzes the Roman prejudice against compulsive gambling (95-101; cf. the story in Calcagnini of the writer Xenophanes, who, when invited to play dice, objected that he was a 'coward in the face of shameful things' (Plutarch De Vit. Pud. 5 = Moralia 530E); and alpha 360, alpha 1899, alpha 2593, alpha 2978, kappa 2592). Serious Romans and Italians, however, objected that gambling in moderation was welcome and useful relaxation from work and war (Calcagnini 287, citing Cicero on Q. Mucius Scaevola in de Oratore 2 and Augustus from Suetonius, Augustus 71; cf. Castiglione, loc.cit.).
A simpler game played by soldiers on a round stone board with 12 playing holes and three pessoi, which might also have served as dice, is known from archaeology (see tau 7 note [5]) but remains unexplained. See Ovid, Tristia 2.481-2 (but S.G. Owen's note in his 1924 edition seems in error), Ars Amatoria 3.365-6 and Isidore, Origines 17.60ff.
[1] Accusative plural(s), the headword one evidently quoted from somewhere (perhaps Euripides, Medea 68; cf. the scholia there).
[2] The reference to Herodotus is obscure, but, as Adler suggests, probably refers to Herodotus 1.94.4, where he describes how the Lydians invented many games (including 6-sided dice and 4-sided knucklebones -- but not pessoi) during a famine in the time of Atys.