Top 5 Computer vs. Human Game Matchups

Video Game System Image Gallery Computer chess is a classic example of a computer vs human matchup. See more video game system pictures.

The story of man versus machine has been popular since the industrial revolution. We humans tend to pride ourselves on our cleverness as a species. No small part of our ingenuity has been dedicated toward designing devices that can help us accomplish tasks more efficiently. In the process, we've raised more questions about whether machines are superior to man.

There's no question that computers are capable of making billions of complex calculations in a fraction of the time it would take most of us to solve just one. Computers can store and access more information on a hard drive than you'd find in an entire library. From a numbers-crunching perspective, computers come out on top.


The challenge for computer programmers is to solve games as if they were math problems. That involves determining the perfect way to play the game so that there are no errors on the part of the computer. With some games, you'll always win if you're the first player to make a move and you follow perfect play. With other games, you may not win but you'll never do worse than a draw.

The amount of time and effort it takes to solve a game depends upon the complexity of the game itself. Computer scientists have solved some games but others remain elusive. We'll look at five matchups between humans and computers, listed in no particular order, that illustrate how far computer science has advanced over the last few decades.

5: BKG 9.8 vs. Luigi Villa

Backgammon introduces an element of chance with dice rolls but that hasn't stopped computers from beating humans.

In June 1979, computer programmer -- and chess player -- Hans J. Berliner watched his backgammon-playing program defeat world champion Luigi Villa by a score of 7-1. It was a remarkable victory. For the first time, a computer program had defeated a human champion at a board game.

Backgammon is a game of strategy and chance. A roll of the dice can convert imminent defeat into victory. That's what happened between Villa and BKG 9.8. Players who analyzed the games said that Villa was the better player but BKG 9.8 benefited from several lucky dice rolls.


Still, the victory marked a turning point in computer intelligence. Berliner explained that his program didn't rely upon a database of moves. Instead, it would analyze the position of pieces on the board and assess the risks or benefits of moving each piece before making a decision. Later backgammon programs became even more proficient at playing against human opponents.

The prize for the exhibition game was $5,000. There's no record of how BKG 9.8 spent its winnings.

4: Chinook vs. Marion Tinsley

Checkers has a limited number of moves and now computers have solved the game.

While some people might think of checkers as a dumb cousin of chess, the game requires strategic and tactical prowess. Perhaps none knew that better than Marion Tinsley, the world champion of checkers from 1955 to 1992. Between 1950 and 1992, Tinsley lost only five games. In August 1992, Tinsley agreed to face off against a new opponent called Chinook.

Chinook began as a project in 1989. Led by Jonathan Schaeffer, Robert Lake, Paul Lu and Martin Bryant, this project would span more than a decade as the team tried to solve the game of checkers. The match against Tinsley in 1992 marked an early attempt to match electronic wits against a human champion.


The first series of matches went well for Tinsley. He emerged victorious, defeating Chinook four games to two, with 33 draws. Tinsley relished the challenge and agreed to a rematch in 1994. After several draws, Tinsley withdrew from the match for health reasons and resigned his title as world champion.

Chinook went on to play and defeat other human challengers like checkers Grandmaster Don Lafferty. In 2007, the team announced they had solved the game of checkers -- perfect play on both sides would always result in a draw.

3: Deep Blue vs. Garry Kasparov

Garry Kasparov considers his next move in the 1997 rematch against IBM's Deep Blue.
AP Photo/Adam Nadel

The year 1996 saw one of the best-publicized machine versus man matchups of all time: IBM's Deep Blue versus chess grandmaster Garry Kasparov. Kasparov was no stranger to playing against computer opponents. In 1985, Kasparov took part in an exhibition that saw him play against 32 computers simultaneously. Kasparov emerged victorious in 1985. He'd do so again in 1996 against Deep Blue.

The match in 1996 consisted of six games. Deep Blue won the first game of the match. Kasparov fired back and won the second game. Games three and four were both draws. Kasparov beat the machine in games five and six.


A year later, Kasparov would meet Deep Blue for a rematch. The new version of Deep Blue was much more powerful. Kasparov won the first game of the match. Deep Blue won the second game. Games three through five were all draws. In the final game of the match, Deep Blue defeated Kasparov and became the first computer to defeat a world champion chess player. Kasparov would request another rematch but IBM retired the Deep Blue project.

Since then, chess-playing computers have become even more proficient. According to the Federation Internationale des Echecs (FIDE) rating system, Garry Kasparov holds the highest rating for human players at 2,851 [source: World Chess Championship]. But now, a chess computer program called Rybka has an estimated rating of more than 3,000 points depending upon the hardware supporting it [source: CHESS].

2: Quackle vs. David Boys

It was 2007 in Toronto, Canada, when a computer program called Quackle beat former Scrabble world champion David Boys in a set of five matches. The minds behind Quackle included another top Scrabble player, Jason Katz-Brown. Quackle is an open-source Scrabble program and is available for download on the Web.

To get to Boys, the Quackle team had to first qualify in preliminary tournaments. Another Scrabble program called Maven competed with Quackle. The program that registered the best ratio of wins to losses earned the right to challenge Boys. Quackle came out ahead.


By building a vast database of words, computer programmers can create viciously effective synthetic opponents. The computer programs scan the board for potential words, sometimes linking letters in unexpected ways. David Boys reportedly shrugged off the loss, proclaiming that it's still better to be human than a machine [source: The Chronicle of Higher Education].

1: MoGoTW vs. Catalin Taranu

Go is a complex game, particularly when you play it on the full 19-by-19 grid.

Go is a game that computers have trouble with traditionally. Played on a board with a grid of either nine by nine or 19 by 19 lines, the game pits two players against each other in an attempt to control territory using black and white stones. The player holding the black stones goes first, placing a stone at an intersection of two lines on the grid. The player holding the white stones goes next. The two players attempt to control territory on the board by surrounding their opponents stones with their own.

The game is tough for computers to play. Unlike games like chess and checkers, Go gets more complicated the longer the game lasts. In chess and checkers, you remove pieces during play as they're captured. In Go, you're placing more stones on the board until you move into an endgame. For this and other reasons, computer programs have had trouble keeping up with human players.


But in July 2010, the MoGoTW program running on 512 cores of the Cray XT4/XT5 supercomputer defeated professional Go player Catalin Taranu in a 19 by 19 Go match. The computer had a seven-stone handicap and won by just 1.5 points.

While Taranu's loss may mark a new era in computer dominance in gaming, it might be premature to dismiss humans. We're clever creatures -- we're not ready to say "game over" just yet.

Lots More Information

Related Articles

  • American Go Association. "What is the Game of Go?" 2010. (Oct. 25, 2010)
  • European Go Congress 2010. "News." July 28, 2010. (Oct. 25, 2010)
  • Fietz, Harald. "Beyond the 3000 Elo barrier." CHESS. May 2007 (Oct. 25, 2010)
  • Gaylord, Chris. "Computers master the game board." The Christian Science Monitor. Aug. 8, 2007. (Oct. 25, 2010)
  • IBM. "Kasparov vs. Deep Blue: the rematch." 1997. (Oct. 25, 2010)
  • Kasparov, Garry. "The Chess Master and the Computer." The New York Review of Books. Feb. 11, 2010. (Oct. 25, 2010)
  • Minkel, JR. "Computers Solve Checkers - It's a Draw." Scientific American. July 19, 2007. (Oct. 25, 2010)
  • Nuttall, Hunter. "Perfect Play: Man Vs. Machine In Games." Jan. 21, 2010. (Oct. 25, 2010)
  • Project Chinook. (Oct 25, 2010)
  • Read, Brock. "A Computer Program Wins Its First Scrabble Tournament." The Chronicle of Higher Education. Jan. 26, 2007. (Oct. 25, 2010)
  • Richards, Mark and Amir, Eyal. "Opponent Modeling in Scrabble." University of Illinois at Urbana-Champaign. International Joint Conferences on Artificial Intelligence. 2007. (Oct. 25, 2010)
  • Sackley, Kristen. "Winning computer program created by graduate student beats world champion Scrabble player." Daily Illini. Feb. 28, 2007. (Oct. 25, 2010)
  • Schaeffer, Jonathan et al. "Chinook: The World Man-Machine Checkers Champion." AI Magazine. Vol. 17. No. 1. Spring 1996.
  • Science News. "Teaching a machine the shades of gray." Science News. Jan. 17, 1981. Vol. 119. Issue 3. pp. 38-39.
  • The Globe and Mail. "Computer backgammon." Sep. 22, 1980. pg. P19.
  • Weeks, Mark. "Braingames 2000 (and more)." World Chess Championship. 2010. (Oct. 25, 2010)