Deep Blue

From Chessprogramming wiki
Jump to: navigation, search

Home * Engines * Deep Blue

Deep Blue [1]

Deep Blue,
the IBM sponsored successor of the chess entity Deep Thought. The project initially started in 1985 as ChipTest at Carnegie Mellon University by the computer science doctoral students Feng-hsiung Hsu and Thomas Anantharaman. Murray Campbell, former co-developer of HiTech, joined the ChipTest team a few months later. The program was named Deep Thought after the fictional computer of the same name from The Hitchhiker's Guide to the Galaxy. Hsu and Campbell joined IBM in 1989, Deep Blue was developed out of this. The name is a play on Deep Thought and Big Blue, IBM's nickname. The declared target was to become the strongest chess entity ever and to beat the human world champion, which eventually happened in 1997 versus Garry Kasparov, winning the $100,000 Fredkin prize, awarded at the AAAI Conference in Providence, Rhode Island.


Deep blue chip [2]


Description given in 1995 from the ICGA site [3] :

Deep Blue Prototype consists of an IBM RS/6000 workstation with 14 chess search engines as slave processors. Each processor contains a VLSI chip for move generation, as well as additional hardware for search and evaluation. Each Deep Thought 2 processor searches about 500,000 positions per second standalone, or about 400,000 positions per second as a slave processor. (This is about 1/10th of the projected speed of the Deep Blue single-processor currently in fabrication.) The 14-processor Deep Thought 2 typically searches between 3 and 5 million positions per second. When conducting a search, the search tree near the root position is processed on the host workstation, and includes selective search extension algorithms such as singular extensions. The deepest nodes in the search tree are handled by the slave search engines which usually do 4-ply alpha-beta searches. 


The 1997 Deep Blue system was based on an IBM RS/6000 SP supercomputer, using 30 workstation nodes of PowerPC processors controlling 16 chess chips each, distributed over two Micro Channel boards. A chess chip features of a full-fledged chess machine on its own, along with move generator, a smart move stack, hardware evaluation function, and an alpha-beta hardware search controller. The search occurs in parallel on two levels, one distributed over the IBM RS/6000 SP switching network and the other over the Micro Channel bus inside a workstation node. A master workstation node first starts the software search exclusively, to distribute work to all 30 workstation nodes only at a certain depth (i.e. 4 plies). After generation an appropriate number of childs and grandchilds etc., the software search per node utilize the 16 chess chips to search the final four plies plus quiescence search in hardware. Since each chess chip could search 2 to 2.5 million nodes per second, the system speed reached about one billion nps (480 chips). During the 1997 match, the software search extended the search to about 40 plies along the forcing lines, even though the nonextended search reached only about 12 plies [4].

Tournaments & Matches

WCCC 1995

Deep Blue Prototype missed the expected win at the WCCC 1995 by losing the decisive match in round 5 against Fritz after king castling into Fritz's half open g-file.

Kasparov versus Deep Blue

Main article: Kasparov versus Deep Blue 1996

Deep Blue was the first machine to win a chess game against a reigning world champion Garry Kasparov under regular time controls. This first win occurred on February 10, 1996, Game 1. However, Kasparov won three games and drew two of the following games, beating Deep Blue by a score of 4–2.

The Rematch

Main article: Kasparov versus Deep Blue 1997

In 1997 Deep Blue won the rematch against Kasparov. He did not recover after the shock by Deep Blues' play in game 2. Kasparov resigned a possibly drawn position, since he missed a deep tricky perpetual check, while he wrongly was confident the machine would not have blundered to allow him to draw. In the final decisive game 6 Kasparov was rather indisposed and blundered in the early opening.

The Deep Blue Team

  • Feng-hsiung Hsu - The man who started the Deep Blue project while still in college
  • Murray Campbell - A former chess champion who works with Deep Blue's evaluation function
  • A. Joseph Hoane - Deep Blue's software engineer
  • C. J. Tan - Senior manager of the Deep Blue development team
  • Jerry Brody - The project's support engineer
  • Joel Benjamin - development team chess consultant, opening book author


5-0 and 5-4.Campbell Benjamin.1997.L02645333.IBM ARCHIVES.lg.jpg

Deep Blue's core team, Joe Hoane, Feng-hsiung Hsu, and Murray Campbell [5]

See also

Selected Publications

1995 ...



AAAI Workshop

  1. Richard Korf (1997). Does Deep Blue use AI? Deep Blue versus Kasparov, AAAI Workshop, pdf
  2. Robert Levinson, Jeff Wilkinson (1997). Deep Blue is Still an Infant. Deep Blue versus Kasparov, AAAI Workshop
  3. Dennis DeCoste (1997). The Future of Chess-Playing Technologies and the Significance of Kasparov Versus deep Blue. Deep Blue versus Kasparov, AAAI Workshop
  4. Selmer Bringsjord, Adam Lally (1997). Chess Isn't Tough Enough: Better Games for Mind-Machine Competition. Deep Blue versus Kasparov, AAAI Workshop, pdf
  5. Fernand Gobet (1997). Can Deep Blue™ make us happy? Reflections on human and artificial expertise. Deep Blue versus Kasparov, AAAI Workshop, pdf
  6. Tony Marsland (1997). The Anatomy of Chess Programs. Deep Blue versus Kasparov, AAAI Workshop, pdf (95)
  7. Carl W. Turner (1997). Attributing Intelligence to Humans and Machines: Between the Devil and the Deep Blue, See? Deep Blue versus Kasparov, AAAI Workshop, pdf
  8. Tony Marsland, Yngvi Björnsson. (1997). From MiniMax to Manhattan. Deep Blue versus Kasparov, AAAI Workshop, pdf
  9. Amol Dattatraya Mali, Amitabha Mukerjee (1997). Modularity Assumptions in Situated Agency. Deep Blue versus Kasparov, AAAI Workshop
  10. Cyrus F. Nournai (1997). Multiagent Chess Games. Deep Blue versus Kasparov, AAAI Workshop, pdf
  11. Kenneth M. Ford, Patrick J. Hayes (1997). What's Wrong With Hal? Deep Blue versus Kasparov, AAAI Workshop, pdf
  12. Franz-Günter Winkler, Johannes Fürnkranz (1997). On Effort in AI Research: A Description Along Two Dimensions. Deep Blue versus Kasparov, AAAI Workshop
  13. David Heath, Derek Allum (1997). The Historical Development of Computer Chess and its Impact on Artificial Intelligence. Deep Blue versus Kasparov, AAAI Workshop, pdf

1998 ...

2000 ...


2010 ...


Forum Posts

1992 ...

1995 ...



2000 ...

2005 ...

2010 ...

2015 ...

External Links

By Date


2000 ...

2010 ...



  1. Deep Blue, a computer similar to this one defeated Garry Kasparov in May 1997. Photo taken by James the photographer at The Computer History Museum, June 14, 2007, Wikimedia Commons
  2. Deep blue chip, International Business Machines (IBM), 1997, Gift of Feng-hsiung Hsu, The Computer History Museum
  3. Deep Thought (Chess) (ICGA Tournaments)
  4. Feng-hsiung Hsu (1999). IBM’s Deep Blue Chess Grandmaster Chips. IEEE Micro, Vol. 19, No. 2, pdf
  5. The core Deep Blue team from The Computer History Museum, Courtesy of IBM Archives
  6. Hans Berliner (1998). Review of Monty Newborn: Kasparov versus Deep Blue. pdf
  7. Re: ICCA Journal Sinks To A New Low by Amir Ban, CCC, January 25, 1998
  8. AAAI Workshop: Deep Blue Versus Kasparov: The Significance for Artificial Intelligence 1997
  9. "Deep Blue" inspires deep thinking about artificial intelligence by computer scientist by Robert Irion, May 5, 1997
  10. Monroe Newborn (1997). Kasparov versus Deep Blue: Computer Chess Comes of Age. Springer
  11. Article by Tesauro on Deep Blue's evaluation-tuning by Stuart Cracraft, CCC, July 11, 2004
  12. New CC book: Beyond Deep Blue by Steven Edwards, CCC, November 11, 2011
  13. Garry Kasparov: Don't fear intelligent machines. Work with them | TED Talk, April 2017
  14. The Day the Machines Took Over by Adam Thierer, Medium, May 11, 2017
  15. Kasparov on Deep Learning in chess by Frederic Friedel, ChessBase News, December 13, 2017
  16. Gerald Tesauro (2001). Comparison Training of Chess Evaluation Functions. In Johannes Fürnkranz, Miroslav Kubat (eds.) (2001). Machines that learn to play games, 117–130, Nova Science Publishers
  17. Strategic Intensity - A Conversation with World Chess Champion Garry Kasparov by Diane L. Coutu, Harvard Business Review, April 2005
  18. 20 years ago, a computer first beat a chess world champion - 1996-1997 The Kasparov-Deep Blue chess matches by Alex Q. Arbuckle, February 10, 2016
  19. "Techmate" by Garry Kasparov by José Antônio Fabiano Mendes, CCC, May 02, 2000
  20. Komodo 8: Deep Blue revisited (1/2) by Graham Banks, CCC, December 27, 2014
  21. Garry Kasparov, Mig Greengard (2017). Deep Thinking: Where Machine Intelligence Ends and Human Creativity Begins. PublicAffairs or John Murray
  22. Scientific American article on Computer Chess by Mark Lefler, CCC, June 03, 2017
  23. 4 hours video: B Larsen met Deep Blue in 1993 in Copenhagen by Jens Bæk Nielsen, CCC, June 01, 2014

Up one Level