Chaturanga

What this variant is, and is not

The base is shatranj, the mediaeval Persian and Arab game, which is well documented. Working backwards from a documented descendant is the standard method, and it is what every published reconstruction does. The board is the ordinary eight by eight, and the men start on the same squares as in shatranj and modern chess.

The defaults this site plays

• The raja, the king, moves one square in any direction, and there is no castling.
• The counsellor beside him moves exactly one square diagonally, like the shatranj ferz.
• The elephant leaps exactly two squares diagonally, jumping over whatever stands between.
• The horse moves as the modern knight, and the chariot as the modern rook. Neither is disputed.
• Pawns step one square forward and capture one square diagonally forward. There is no two-square advance and no capture in passing.
• A pawn reaching the far rank promotes to a counsellor, and to nothing else.
• The raja may make one knight leap in the whole game, provided he is not in check at the time.

The three disputed points

Each of these is a genuine scholarly disagreement, not a house rule. Where a game offers the alternative as a setting, it is named here.

The elephant’s move

The default here is a leap of exactly two squares diagonally, which is Murray’s reading and follows the Arabic sources directly. Bock-Raming’s work on the Sanskrit material argues instead for a leap of exactly two squares orthogonally, along a rank or a file. That alternative is available as a setting. The two elephants are quite different pieces in practice, and the choice changes the whole shape of a game.

The king’s knight leap

Murray suggested that the raja originally had a single knight-style leap, usable once per game, and that it is the ancestor of castling. This site switches it on by default, restricted so that it may not be used to escape check. It can also be switched off, which leaves the raja moving exactly like a shatranj shah.

Stalemate

The sources point in three directions, so this site takes the neutral option and scores stalemate as a draw. The two alternatives are both defensible: the player who stalemates wins, which is the shatranj rule, or the stalemated player wins, which is recorded in some later Indian practice. Both are offered as settings.

Winning

• Checkmate. The enemy raja is in check with no legal answer.
• Baring the enemy king: capturing every enemy man except the raja. As in shatranj, if the bared player can bare your king in reply on their very next move, the game is a draw instead.
• Resignation, or the opponent running out of time.

Drawing

• Stalemate, under this site’s default setting.
• Agreement between the players.
• Threefold repetition, and the fifty-move rule, both applied automatically for both sides.
• A bare king answered immediately by baring the opponent’s king.

Sources

Everything on this page is written from scratch. These are the works it leans on, and what each one settles.

• H. J. R. Murray, A History of Chess (Oxford, 1913)

  The source of this site’s defaults: the two-square diagonal elephant, following the Arabic evidence, and the suggestion that the king once had a single knight leap. Murray is also the reason the reconstruction is built outward from shatranj rather than from the Indian texts alone.

• Andreas Bock-Raming, studies of the Sanskrit and Indian chess sources

  Close reading of the Indian material, which argues that the elephant moved two squares orthogonally rather than diagonally. That reading is the alternative setting offered here, and it is the main reason this page treats the elephant as unsettled.