5

Kasparov said, in Game Over: Kasparov and the Machine, that Deep Blue "sank into deep thinking" and took like 15 minutes to make a move.

The movie tells the story of their match, in 1997.

At a certain position, in the second game, Deep Blue could have taken two pawns, but did not, aiming for a win in the long run.

Do modern computers agree with Deep Blue on that, and, if they do, how long does it take them to find the correct winning move?

That part of the movie can be watched here.

  • This is the second game of the 1997 match: chessgames.com/perl/chessgame?gid=1070913 Is this the game you are referring to? Deep Blue did not give up any material in this game. – Ken Wei Aug 7 '15 at 12:00
  • 2
    I see a question in the title and a question in the actual text, but I see no connection between them. Which one is the question you want to ask? – JiK Aug 7 '15 at 14:20
  • @JiK Well, the title is a concise description of what I'm asking. The text explains further what I meant. They should be seen as complementary, as most texts and their titles are. Also, did you miss the third question? – Daniel Aug 7 '15 at 18:19
  • Basically imo Kasparov just made up lots of stuff about Deep Blue playing human moves on the basis that computers didn't normally play like that. Without realising that it is subroutines which run computer programs, which may be vastly different. It's like saying Windows 95 is suspicious as an operating system because it's nothing like Windows 3.1. – magd Aug 7 '15 at 18:34
  • @magd Yes, but there are some other points involved. By "sank into deep thinking", Kasparov implies the computer never spent that long time on any other move. It was also said that the computer "should" capitalize on the pawns, instead of doing what it did. He said that someone might have told the computer to drop the original "grab the free pawns" line. – Daniel Aug 7 '15 at 20:23
6

This article from Chessbase answers most of your questions.

This was the position in which Kasparov had set a "computer chess" trap.

 [fen "r1r1q1k1/6p1/p2b1p1p/1p1PpP2/PPp5/2P4P/R1B2QP1/R5K1 w - - 0 35"]
[White "Deep Blue"]
[Black "Gary Kasparov"]

He was expecting 36. Qb6 Rd8 37. ab Rab8 38. Qxa6 e4

 [fen "1r1rq1k1/6p1/Q2b1p1p/1P1P1P2/1Pp1p3/2P4P/R1B3P1/R5K1 w - - 0 35"]
[White "Deep Blue"]
[Black "Gary Kasparov"]

when, for the cost of 2 pawns he has generated dangerous counterplay based on the idea of Qe5 possibly followed by e3.

Instead Deep Blue played the immediate ab followed by Be4 thereby nipping in the bud any chances of counterplay via ... e4 and ... Qe5 for black.

According to the article -

In fact, if you consult any of the top engines of today, whether it be Houdini 4, Stockfish 6, or Komodo 8, they all choose Deep Blue's move 36.axb5. For example:

Komodo 8: 36.axb5 axb5 37.Be4 Qd8 38.Kh2 Rcb8 39.Ra6 Kf8 40.R6a5 Kg8 41.R1a2 Rxa5 42.Rxa5 Bc7 43.Ra1 Bb6

+/- (0.84) Depth: 29 00:02:15 686MN, tb=57

So, nothing to see here (2 min 15 sec for the computer) ;-) Move on!

  • I wasn't aware of that article, of course (they took a while to write about that!) - but I had this question for a while now and thought it would be interesting for the community. There's some good discussion happening there - even Joel Benjamin decided to join in. Even though I think everyone will agree there was no cheating involved, I guess most people won't ever forgive IBM for its lack of touch. Great answer, btw. – Daniel Aug 9 '15 at 21:49
  • 1
    @DanielGomes I think Kasparov's mistake in this match was a very natural one, one which most of us make. Instead of playing the position he became distracted by trying to play the man/computer. When he thought he was playing more man than computer this disturbed his equilibrium to such an extent that he started playing bad moves. It is part of the human condition and one which none of us can escape completely, even Kasparov. – Brian Towers Aug 10 '15 at 6:39

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.