The History of Checkers

Checkers as we play it, on the dark squares of a chessboard, took shape in medieval Europe, most likely in 12th-century France. It probably grew out of alquerque, an older jumping game, and gained its defining rule, compulsory capture, around the 16th century. The English-speaking version got its first book in 1756 and its first world champion in 1840, and in 2007 the team behind the Chinook program proved that perfect play ends in a draw.

The 5,000-Year Claim

Many rule sheets and websites say checkers is thousands of years old and was played first in Egypt or Mesopotamia. The evidence does not support it. The boards found at Ur, from about 3000 BC, belong to a race game played with dice and seven pieces a side, which works nothing like the jumping captures of checkers.

The Egyptian case is weaker still. Carvings at the temple of Kurna, often dated to 1400 BC, include Coptic crosses, which point to a much later date, and have been read as three different games. Greek petteia and Roman latrunculi were capture games on squared boards, alike only in that pieces could be captured.

Alquerque, the Likely Ancestor

The best candidate for an ancestor is alquerque, played on a grid of 5 x 5 lines with 12 pieces a side on 25 points. Pieces moved along the lines to a neighbouring point and captured by jumping, sometimes several times in one turn. The game first appears as qirkat in the 10th-century Kitab al-Aghani of Abu al-Faraj al-Isfahani, which names it without giving rules.

The Moors carried it to Spain, and in 1283 the Libro de los juegos, the book of games commissioned by Alfonso X of Castile, set down how to play it. The book does not make clear whether capturing was compulsory, and it never mentions promotion. The jumping capture is the clear link. Otherwise the two games differ: the oldest alquerque rules have no crowning, and alquerque starts with nearly every point filled, while checkers leaves two empty rows between the armies.

Fierges: Onto the Chessboard

At some point a jumping game moved onto the chessboard and began using only the diagonals. Most accounts put this in France, or perhaps Spain, in the 12th century. The French called the game fierges, the name of the chess queen of the time, which moved just one square diagonally, much like a checkers king.

When the chess queen was renamed dame, the checkers pieces took the same name, which is why the game is still called dames, damas, or dama in much of the world. Using only the diagonals wastes half the squares, but men that can only step forward can never shuffle back and forth, so every game moves toward an end.

Compulsory Capture and the Huff

Early checkers let players decline a capture. Compulsory capture arrived in France in the 16th century, and for a time two games existed side by side, one where you had to take and one where you did not. The forced game is the closer of the two to modern American checkers.

Players enforced the new rule with the huff: miss a capture and your opponent could remove the piece that should have jumped. That made capturing optional in practice, since a player could accept the huff when it cost less than the capture. European play dropped the huff in the late 19th century, English-speaking countries last of all, and today's official rules simply treat a missed capture as an illegal move.

The 10x10 Game Splits Off

In the early 18th century a larger game spread on a 10x10 board. It was called Polish draughts, though it probably did not come from Poland. It gave kings the power to fly along the diagonals, let men capture backward, and made players take the most pieces available. As international draughts it became the main game in France and the Netherlands, held its first world championship in 1885, and gained a world federation in 1947, founded by France, the Netherlands, Belgium, and Switzerland. The British and American game kept the older short kings.

Checkers in Britain and America

The first book on the game in English was William Payne's An Introduction to the Game of Draughts, published in 1756, with a dedication and preface by Samuel Johnson. Serious competition followed in the 19th century. Andrew Anderson of Scotland became the first recognised world champion in 1840, and his countryman James Wyllie held the title in four separate spells between 1844 and 1894.

From 1840 to 1994 every men's world champion came from Scotland, England, or the United States. American championships used a two-move opening restriction from 1900 and switched in 1934 to the three-move ballot, which draws the first three moves at random and cuts down draws between experts. Since 1994 the title has also gone to players from Barbados, South Africa, and Italy.

Marion Tinsley

Marion Tinsley (1927-1995), born in Ironton, Ohio, is widely rated the greatest checkers player of all time. A professor of mathematics at Florida State University and Florida A&M University, he was world champion from 1955 to 1958 and again from 1975 to 1991, and he never lost a world championship match.

By the usual count he lost only seven games from 1950 until his death in 1995, and two of those were to a computer. In 1992 he played the program Chinook for the man-versus-machine title and won with 4 wins, 2 losses, and 33 draws.

Chinook and the Solved Game

Checkers was an early test bed for computer game playing. Christopher Strachey wrote a draughts program in 1951, and in the 1950s Arthur Samuel at IBM wrote a checkers program that improved by playing against versions of itself, one of the first programs that learned.

Chinook was started in 1989 at the University of Alberta by a team led by Jonathan Schaeffer. In the 1994 rematch with Tinsley, the first six games were drawn before Tinsley withdrew with pancreatic cancer, and Chinook became the first program to win a human world title. Tinsley died in April 1995. That year Chinook defended the title against Don Lafferty, 1 win and 31 draws.

Schaeffer's team then set out to solve the game. In July 2007 they published the proof in the journal Science: from the starting position, perfect play by both sides is a draw. It took 18 years, with up to 200 desktop computers working at the peak. Checkers has about 500 billion billion possible positions, and the proof covers the normal start, not all 156 openings of the tournament ballot.