Skip to main content
Search IBM Research
     Home  |  Products & services  |  Support & downloads  |  My account
 Select a country
 IBM Research Home
Deep Blue
The Match
 ·Post-Match Images
 ·Last year's match
 ·The rules of the rematch
 ·The Commentators
The Players
The Technology
The Community

Related Links
 Press room
 Chess conference
 Site guide
 Search Research

Deep Blue game 6: May 11 @ 3:00PM EDT | 19:00PM GMT        kasparov 2.5 deep blue 3.5

White: Deep Blue
Black: Kasparov
1. e4 e5
2. Nf3 Nc6
3. Bb5 a6
4. Ba4 Nf6
5. 0-0 Be7
6. Re1 b5
7. Bb3 d6
8. c3 0-0
9. h3 h6
10. d4 Re8
11. Nbd2 Bf8
12. Nf1 Bd7
13. Ng3 Na5
14. Bc2 c5
15. b3 Nc6
16. d5 Ne7
17. Be3 Ng6
18. Qd2 Nh7
19. a4 Nh4
20. Nxh4 Qxh4
21. Qe2 Qd8
22. b4 Qc7
23. Rec1 c4
24. Ra3 Rec8
25. Rca1 Qd8
26. f4 Nf6
27. fxe5 dxe5
28. Qf1 Ne8
29. Qf2 Nd6
30. Bb6 Qe8
31. R3a2 Be7
32. Bc5 Bf8
33. Nf5 Bxf5
34. exf5 f6
35. Bxd6 Bxd6
36. axb5 axb5
37. Be4 Rxa2
38. Qxa2 Qd7
39. Qa7 Rc7
40. Qb6 Rb7
41. Ra8+ Kf7
42. Qa6 Qc7
43. Qc6 Qb6+
44. Kf1 Rb8
45. Ra6 1-0

Game 2:

So this proves it, last year's win for Deep Blue was no fluke. The computer stunned the audience and particularly the watching Grandmasters by playing a seamless strategic game in a type of blocked position that conventional wisdom has always held, favors the human player.

Maybe its time to rewrite the textbooks, the match is alive again at 1-1. Kasparov was stunned by his defeat and left the playing area at great speed without comment. The Deep Blue team received a standing ovation and their chess expert Joel Benjamin summed it up: " This was real chess."

Kasparov has the advantage of the white pieces in Tuesday's third game, and will welcome the rest day to collect his thoughts. Deep Blue has to find an improvement over its play in game 1 but whatever the outcome, its clearly worthy of its nickname Deeper Blue.

Pre-game commentary:

YASSER SEIRAWAN: -- was when this opening was introduced.
stop and think about that for a moment. The 15th century,
they've been playing this hoping ever since. The idea of the
opening is what we call the classical center. This is depicted
by the d4 pawn and the e4 pawn.
MAURICE ASHLEY: Another move has been played, Na5-c6.
YASSER SEIRAWAN: White in this opening seeks to maintain --
MAURICE ASHLEY: I guess not, because he just advanced the
to d5, attacking the knight on c6, so clearly we're still in
opening book. This is opening they're /REUFPLT this is the
kind of thing that has been played several times. And Kasparov
quickly /SPOPBGDZ, moving his knight back to e7 from c6.
MAURICE ASHLEY: And Deep Blue responding immediately Be3.
Usually I get to do play-by-play. This is getting a little
YASSER SEIRAWAN: Push, push, push.
MAURICE ASHLEY: And Kasparov playing instantly. We're seeing
we've seen 17th moves in all of two minutes, maybe, it's just
instantaneous chess going on. At some point someone's got to
be out of their book.
MAURICE ASHLEY: And we'll get a chance to breathe and explain
little bit of what's going on. And after Ne7-g6 Kasparov has
gotten up from the chessboard. He has a dressing room in back
of their playing area so that he often retires to that area
particularly when he feels comfortable about the position, and
with his vast knowledge of opening theory as we're seeing him
evidence here, it's clear that he's very, very comfortable,
probably is just where he wants to be, he is not worrying about
everything, and he's expecting Deep Blue to come up with
something new, something interesting.
This last move, Ng6, has caused Deep Blue to hesitate just a bit,
Yaz, he's not played immediately so maybe we've seen the
innovation that Kasparov has sprung so that Deep Blue will have
to think on its own.
MIKE VALVO: It's out of book.
YASSER SEIRAWAN: Indeed. The idea is that had Garry played
of the moves in Deep Blue's computer library, Deep Blue would
have responded quickly. So the move Ng6 is probably out of its
library now, and now Deep Blue is going into its first think of
the game. And let's see if we can try to explain what's going
on in the position.
Well, first of all, what ourselves spoke about yesterday is
what's the ideal way of playing against the computer? And in
the competition at Aegon, a recent competition I played in The
Hague, it was quite clear that what the human player sought to
do was block aid the position, and the definition of a block
aid is where the pawn chains interlock in such a way that all
the lines are closed.
So we have a closed opening here. If you look at this chain of
pawns in the middle of the board, you notice that no captures
by pawns are possible, and so it becomes not a game of play in
the center, but rather, play on the flanks. Whenever you have
a game where the game is being played on the flanks, you're
going to see a great deal of maneuvering. And in this part of
chess, the human players are extremely good. And the --
they're our counterparts, our silicon /POURPBT parts have not
caught up. So computers tend to get out-maneuvered in closed
So what Garry has achieved, it's not his bailiwick, what he would
normally like to do, but he has achieved his closed position.
The situation on the king-side is rather interesting. If you
notice all four knights are rather balanced, both knights --
white may spring his knight to the f5 square with Ng3-f5 or
vice versa black may play Ng6-f4. So we have a very balanced
position on the king-side. The queen-side, not too much is
happening on the queen-side. Probably white at some point in
the future will have to play the move a2-a4 to unlock the
secrets on the queen-side.
MAURICE ASHLEY: Considering that blocked positions are -- that
computers are known to play blocked positions notoriously
poorly, Mike, why do you think that the programmers, and I
guess more specifically Joel Benjamin, would allow this kind of
situation to occur when he knows his baby doesn't do as well
under these conditions.
MIKE VALVO: Well, he can't look at actually everything in his
preparations, and he made a rough assumption that he would not
very likely get into this line, so he just programmed in some
book that was generally accepted that everybody plays and left
it at that. He probably never even looked at this more than
five or ten minutes, I would bet. And --
YASSER SEIRAWAN: And just to buttress that point, I'm sure that
at some point during their 14-month preparation for this match,
that Deep Blue's team ever had this specific position in their
study. And we do have another move.
MIKE VALVO: And indeed, Deep Blue may have never actually
this position. It was all programmed in. It was nothing knew
MAURICE ASHLEY: The last move by Deep Blue, queen from
d1-d2 has
been played and Kasparov has quickly come back to the board.
But still let's follow that point a little bit, since it seems
very interesting to me. How is it that you program moves into
the computer -- I mean 14 months is a long time. What do they
do so that Deep Blue knows that this is the move it should
play? Why did it play so quickly, if these moves were
programmed in but not looked at at all? Who was programming in
the moves and not looking at the moves? Seems to be a
disadvantage for the computer to be --
MIKE VALVO: Yeah, it is. And sometimes all they can do is have
the computer analyze all of the end or terminal positions of a
line and see if it likes it. And if it likes it, they'll
accept it. Sometimes they can do a deep analysis of it, and
sometimes they can't. Sometimes they'll let it think at night
when nobody's watching, and unless there's some kind of a
problem, they'll /KP-RT it. But yes, this is not the ideal
position for a computer. Humanness is great because a human
thinks conceptually and concepts of maneuvering is sort of
required, and computers kind of think a move at a time here.
They're just trying to find the best position. The best
position for them is where they have the most space, the most
lines, the most squares they can go to, and in this kind of
blocked position, it doesn't feel good very well.

Related Information

      Program of events:

      Chess links:

  About IBM  |  Privacy  |  Legal  |  Contact