Ever since chess was created roughly 1500 years ago, it’s seen both grandmasters and casuals alike attempting to understand the game. Recently, I myself have begun learning Chess on a website called Chess.com. Chess.com is able to tell me how well I performed after every one of my matches. But how does it always know which is the best move to play? I felt intrigued to explore the history of chess programs and what their impacts have been on the game.
What is chess?
Chess is a strategic board game, played on an eight-by-eight chessboard. The objective of the game is to checkmate the opponent’s king, trapping it so that it is unable to move without being captured. This is done using five other unique pieces, including two rooks, two knights, two bishops, eight pawns, and a queen. Each player takes turns moving pieces around the chessboard, attempting to win small advantages and eventually trap the opponent’s king.
Chess AI: A brief history
Alan Turing led one of the first endeavours into creating a chess-playing computer back in 1946. Most importantly, he assigned values for each piece in terms of pawns that are still in use today: pawn = 1, knight = 3, bishop = 3.5, rook = 5, queen = 10, and king = 1000.
The first pieces of software for chess took the form of problem-solving programs. The 1951 Ferranti Mark 1 solved predefined “mate-in-two” problems. However, computational power, not code, limited these programs. There are more moves in a game of chess than there are grains of sand on all the beaches in the world! Yet, as computational power and programming techniques evolved, so did chess programs. New programs were able to successfully play and even beat human chess players. Innovations continued until finally in 1997, a computer program called DeepBlue successfully beat the reigning world chess champion, Garry Kasparov. From then on, even more advanced computer programs, capable of calculating the best moves while looking dozens of moves ahead, reigned supreme.
The future of the game
Chess programs continued to advance, building off one another. Each program accessed a huge detailed library of different moves and combinations that the program could analyze. However, the introduction of AlphaZero changed all that. AlphaZero is a self-taught AI chess program, initially given only the rules of Chess. In just four hours of playing itself, AlphaZero became the strongest chess program in the world, beating the reigning chess program StockFish 155-6. It was unpredictable, playing variations that no human could play because of its ability to predict its opponent’s moves far into the future. Yet many players, including the best chess players in the world, now analyze games played between programs to learn tactics and sequences never seen before.
Although computers might dominate the game, AI will never replace the human-to-human, face-to-face showdowns between players. Chess is still a game of many tactics, strategies, and mental tricks between players. However, there is still much to learn from AI that will ultimately benefit the game as a whole.