The Origin of Chess: from the game structure

L'origine des échecs : d'après la structure du jeu


Quoi de neuf ?

Mes livres

Histoire des échecs



With differentiated pieces, captures by substitution and victory by capture of the central piece, Chess does not look like any other boardgame. Two forms seem "primitive", the "indo-persan" and the chinese. They present obvious similarities. Therefore, both games are cousins but the process which would have derived one from the other is still unknown. More than their similarities, their differences may give some indications on their "genealogy". Their compared structures, material and rules, may allow to understand which game precedeed the other.

 Indo-persian Chess, Chatrang or Chaturanga, is played on squares with 8 major pieces and a full line of Pawns which do not move as they capture and which promote when reaching the opposite side of the board.


Chinese Chess, Xiangqi, is played on the intersections of a boad whose marks influence the moves of 11 major pieces, 2 among them being with no occidental equivalent. The Pawn line is incomplete and those Pawns keep a very regular move. (An exhaustive comparison of both Chess game is available here). If the occidental game represents a front head battle, the oriental game is a palace assault, a blocking game.


 From the Mediterranean to India, race games are numerous and known from Antiquity. All of them have their men moving upon squares. Also, the 64 square board was used for Ashtâpada, a dice game used since 3rd century BC at least in India, and maybe in Persia (Hashtpay, Panaino) as well. In China, predecessor games were Weiqi and Liubo, possibly both evoked in Confucius' Analects around 500 BC. The first one was played (and still played) over intersections and the second one used 6 pawns (including 1 chief) per player, had a board with many marks and had a central square named "water". Liubo declined under Tang dynasty (618-907), then disapeared but a relation with the "river", the 5 Pawns and General of Xiangqi is tempting.

The Liubo board

Some experts have proposed that the move of the major pieces is in relation with magic squares. The central pieces (King or equivalent and his Counsellors) would cover the first ring (around a central square), the other pieces sharing the second ring. That would explain why the Horse (Knight) has such a strange jump. The Elephant finds his way, however the Rook is out of the scheme. Except if the Rook had a different move at the beginning, like a jump to 2nd square on line and column, which is appealing but has never been prooved by any evidence. In overall, the idea behind the magic square theory is good, however, up to now, that theory remains a speculation.

 Moves reached from the central square: by the King, Minister or Counsellor, Elephant, Knight. The greyed squares could correspond to a primitive Catapult (speculative).
Remark: only in Chinese Chess the King is restricted to the 4 orthogonal squares; in Indo-persian Chess (as well in Korean), the King goes also on the M squares.

The Pawn diagonal capture mode in Chatrang is intriguing. The greek game of Petteia with its custodian capture might have brought an influence. Indeed at this game, the target man is located at one diagonal step from the offensive man. It is well known that the hellenic influence lasts several centuries in Central Asia after Alexander The Great's passage. In comparison, the Chinese Pawn looks basic. Does that mean that it is anterior to the adoption of the diagonal capture and the last-row-promotion in the West?

Concerning the King, Xiangqi with a central position looks more primitive than the Indo-persian game which looks more like an evolution than a invention. Who would invent an asymmetrical game where both King's flanks are not balanced?

More generally, the moves of the other Chinese pieces, with forbidden jump, blocked in the palace or by the river, seem more primitive than the free pieces of Chatrang. And it is easier to see the 3D carved indo-persan figurines as an improvement (to make the game more appealing, especially for those not able to read) of the Chinese tokens than the reverse way. The point is that if the main stream opinion believes in a birth in India, nobody has explained yet, in a satisfactorily manner, how Xiangqi could have evolved from Chatrang/Chaturanga.

In contrast, the Chinese Cannon appears modern. But it is known that it was first a Catapult and its first move (before 1206, first description in the Qiguo Xiangxi) is unknown. Some have emitted the idea that it could have been an orthogonal two-steps jump, the only elementary move missing at Chess. Speculation again, but interesting idea. The modern move would then be a more recent change. Amusingly, when Arabs invented enlarged variants on 10x10 board, some gave this particular move to the additional piece and named it Dabbâba. This word means a siege machine in mediaeval Arab. Trebuchet, onager. This is not very far from the Catapult...

Imagine you get a purse with 32 tokens, all of the same size, and you know that this is for playing a battle game that you once saw in China. You get a board, it has 8x8 squares, you start placing the token on two lines per player. You try to remember the moves. You forget about the river and palace limitations because your board, an ashtapada board, does not have these markings. What you have is close to the Western game.

Now, try to imagine the opposite. You get a purse with 32 carved and figurative pieces. You can not forget that the board was a plain one with 64 squares because it is so simple. But you prefer to invent a new one with markings. You want those markings influencing the moves of all pieces. You restrain their freedom. And, miracle, everyone in China follow you, nobody play like in the West, they all play the game you invented.

Choose your scenario.


It is our belief that historians have not paid enough attention to the structure of the games so far. The Chinese game, Xiangqi, appears more primitive than its western counterpart in India or Persia.

Let us imagine that Chess was invented in China in a proto-Xiangqi form. It is easy to understand why it would have evolved into a game with figurine and a simpler board: because Western people did not use the Chinese writing system! But, if we imagine the game was invented as a proto-Chatrang-Chaturanga with figurines and a plain board, it is very difficult to understand why it was not adopted by Chinese. Afterall, Chinese had adopted Nard becoming their Shuanglu with no modification at all (and Shuanglu displaced their Liubo which became extinct) (It is amazing to see that Shuanglu still retains the same starting arrangement than modern Backgammon).

If Chatrang-Chaturanga had been introduced in China, how come it could have evolved into Xiangqi? Nobody has given a satisfactory explanation so far.

In the meantime, I distribute the points:

  • China: 3 points
  • Persia: 1 point
  • India: 1 point

If we only had their structure to decide, there is no question, Xiangqi would be proclamed the ancestor. But the things are not so simple...



 Les arguments présentés sur cette page sont développés dans L'Odyssée des jeux d'échecs, (Praxeo Editions, 2010).