History of Checkers: Alquerque, Draughts and Chinook

The jumping game in Alfonso X’s manuscript

The strongest early evidence is a book you can still see. The Libro de los Juegos, completed in Seville in 1283 under Alfonso X of Castile, illustrates and describes several games called alquerque. The version most often compared with draughts is alquerque de doze, named for the twelve pieces on each side. Its board has twenty-five points arranged in five rows, with lines connecting them. The players begin with the center point empty.

Pieces sit on intersections and travel along the lines. A capture hops an opposing piece into an empty point beyond it, and a piece can take several opponents in a single turn. Those jumps look familiar to a modern player. The geometry is different. Some lines run straight, and play starts almost completely packed, with one empty point to open the game.

Older boards and markings are sometimes identified as ancestors of draughts. A grid scratched into stone can tell us about the shape of a board. The rules people used on it are harder to recover. Alfonso’s manuscript gives historians both a picture and a description of play, which is why it is a firmer starting point than claims about a particular ancient birthplace.

From points on lines to squares on a board

A familiar account places the transition to the 64-square board in medieval France. Iberian evidence has also led historians to propose a Spanish origin for the later game. The surviving records leave the date and place of that transition open. By 1591, Pedro Ruiz Montero had published a book on the game of damas in Valencia, evidence that the Iberian tradition already had enough theory to fill a manual.

The change in board explains a lot about how modern checkers feels. On an 8×8 board, a diagonal move from a dark square always lands on another dark square. That gives the game thirty-two playable squares, with lanes that spread toward the center and narrow along the edges. In English draughts, ordinary pieces go forward. A piece that reaches the far side is crowned and can come back the other way.

Capture customs also varied. Under the old practice of huffing, a player who overlooked a jump could lose the piece that should have made it. Under modern English draughts rules, the jump itself is compulsory. When a player has a choice of captures, they may choose their route, and then must continue the chain with that piece until the turn ends.

Printed games give players a way to study

In 1756, London mathematics teacher William Payne published An Introduction to the Game of Draughts. It collected fifty games and selected positions that readers could set up and work through. A player could now learn from a strong line of play even without a strong opponent in the room. Joshua Sturges’s guide followed in 1800. An expanded Brooklyn edition appeared in 1881, an example of British draughts literature finding a second life among American checkers readers.

The record of a game could fit on a small page because the playable squares were numbered from 1 to 32. Once readers knew the numbers, they could follow a move such as 11–15 without a full drawing of the board. A book could publish the same opening with several possible replies and show exactly where the game took a different course. Names such as the Old Fourteenth gave players a shared way to refer to a line without reciting every move.

American players added a literature of their own. Henry Spayth published The American Draught Player in 1860, and William Timothy Call’s Vocabulary of Checkers appeared in 1909. Call’s dictionary records the language of problem composers, opening specialists and match players. The game had acquired a technical vocabulary even though it still needed only a board and two sets of discs.

Championships change the first moves

In nineteenth-century match play, rivals could face each other across many games. Call identifies the 1864 match between James Wyllie and Robert Martins as an important early example of restricting the opening. Match records show the custom evolving. The 1894 championship between James Ferrie and Wyllie included games with free opening choice and games with a two-move restriction.

There was a practical reason to vary the start. If both experts knew a dependable opening, they could repeat the same well-studied line and make a long match revolve around a few familiar positions. A ballot puts the opening on a card. In three-move play, the card gives the first move by the opening side, the reply, and the opening side’s next move. From there, players have to make their own decisions.

The older format survives as go-as-you-please, often abbreviated GAYP. The distinction is in who chooses the opening. The moves, jumps and crowning that follow work the same way. LarkSpark games start from the regular position and leave the opening choices to the players.

One family of games, several boards

English draughts and American checkers are two names for the 8×8 version covered in this guide. The international game took a different path. An eighteenth-century Parisian manual by Manoury describes the 100-square game then called Polish draughts. Even that name leaves the origin open. The French federation points to an earlier dated 100-square board in the Netherlands, among the clues that complicate the Paris origin story.

Today, international draughts uses a 10×10 board with twenty men on each side. Its ordinary pieces can capture backward, its flying kings can travel along an open diagonal, and a player must choose a route that takes the greatest number of pieces. English draughts keeps short-range kings and allows any legal capture route. A player who switches between the two has to relearn which pieces are safe, even though the board and discs look familiar.

Checkers becomes a test for machine learning

At IBM in the 1950s, Arthur Samuel used checkers to study how a program could improve at a task with experience. He published the work in 1959. His program looked ahead through possible moves and scored positions using features it could calculate. One learning method remembered positions it had already analyzed. Another adjusted how it evaluated positions as it gained experience.

Checkers made a good experiment because a program could work with a precise set of rules and still face difficult choices. For example, counting pieces alone can make a sacrifice look like a blunder, even if it forces a double jump on the next turn. The program had to search far enough ahead and judge the positions it found along the way.

A world champion faces a computer

Decades later, a University of Alberta team led by Jonathan Schaeffer built Chinook. It finished second behind world champion Marion Tinsley at the 1990 U.S. championship without losing a game, earning a chance to challenge him.

The American Checker Federation and English Draughts Association initially refused to let a computer play for the human title. Tinsley resigned the title to play Chinook anyway, and the two associations created a separate man–machine championship. In London in 1992, Chinook won two games and drew thirty-three. Tinsley won four and took the match.

They met again in 1994. The first six games were draws. Tinsley withdrew because of ill health, and Chinook took the man–machine title by default. The match left a question the team could answer by computation. How well could a program do against an opponent who always found the best move?

What it means to solve checkers

In 2007, Schaeffer and his colleagues published the answer in Science. From the standard starting position, perfect play by both sides produces a draw. The team estimated that checkers has about 5 × 10²⁰ possible positions. They proved the result by combining search from the starting position with endgame databases, where positions with ten pieces or fewer had already been evaluated as wins, losses or draws.

This is called a weak solution. It establishes the value of the initial position and gives each side a strategy that guarantees at least a draw from there. A solution for every legal position would be a stronger result. The proof still leaves plenty for players to discover. A single missed jump can turn an even game into a loss, and deciding which of two similar-looking king endings can be saved is the kind of work that kept players publishing new analyses for centuries.

The History of Checkers