Author: Uri Blass
Date: 01:21:13 02/02/03
Go up one level in this thread
On February 01, 2003 at 20:14:09, Amir Ban wrote:
>[Event "X3D Man-Machine match"]
>[Site "New York City"]
>[Date "2003.01.30"]
>[Round "3"]
>[White "Kasparov, Garry"]
>[Black "Deep Junior 8"]
>[Result "0-1"]
>[Opening "QGD semi-Slav: Stoltz variation"]
>[ECO "D45"]
>
>1. d4
>
>{I was operating again, and this time the guest of honour was legendary director
>Milos Forman. He made this move for white, then sat down in the VIP section of
>the playing room for the rest of the game. There are about 10 chairs for VIP's
>about four meters from the board, but they are usually unoccupied because it's
>much more interesting three floors down where Ashley and Seirawan are commenting
>for the crowd. Klara Kasparova, however, is always there, keeping an eye on her
>son}
>
>1... d5 2. c4 c6 3. Nc3 Nf6 4. e3 e6 5. Nf3 Nbd7 6. Qc2 b6
>
>{Here's where we differ from the previous game's 6...Bd6. Are we now going to
>see a conventional line from the Semi-Slav ? Not if Kasparov can help it}
>
>7. cxd5 exd5 8. Bd3 Be7
>
>{We are out of book here}
>
>9. Bd2
>
>{What's this ? Is he going to castle queenside ?}
>
>9... O-O 10. g4
>
>{So that's the idea again. This time, we have to take}
>
>10... Nxg4
>
>{Deep Junior expects 11. Bxh7+ Kh8 12. h3 Ngf6. While Kasparov was considering
>his next move I went out to the computer room and found Boris, much troubled,
>analysing continuations after this line. I asked him: "Will he take on h7?". "He
>has to", says Boris. "Can't he play say Rg1 ?". "No, then I play Ndf6" says
>Boris.
>
>11. Rg1
>
>{Kasparov is not listening to Boris. Maybe in the spirit of the g4 move, white
>doesn't want the pawn back, but much more. I believe he was out not only to win,
>but to completely humiliate the program and get it over by move 15. This may yet
>happen on another day, but today Junior found the right moves}
>
>11... Ndf6 12. h3 Nh6 13. e4 dxe4
>
>{This looks very dangerous now, but Deep Junior is not concerned. Evaluations
>are level, and not falling. Every iteration that doesn't bring a massive
>fail-low is a relief}
>
>14. Bxh6
>
>{We expected 14. Bxe4 Kh8 15. Bxh6 gxh6 16. Bxc6 Rb8 17. Qd2 Bxh3 18. Qxh6 Bg4
>with advantage for white}
>
>14... exd3 15. Rxg7+ Kh8 16. Qxd3 Rg8 17. Rxg8+ Nxg8 18. Bf4 f6
>
>{Deep Junior didn't like 18... Qd7 19. Ne5 Qxh3 20. Qxh3 Bxh3 21. Nxc6 Bh4 22.
>d5 and instead opted for this key defensive move, shutting out e5 and g5 for
>white's minors. With the knight at g8 the king looks well protected now. At
>around 0.5, this was the most negative score in the game, but in the next few
>moves Deep Junior believes it equalizes}
>
>19. O-O-O Bd6 20. Qe3 Bxf4 21. Qxf4 Bxh3
>
>{This looks safe now. It is rare to see Junior defending a pawn up. It goes
>against its temperament}
>
>22. Rg1 Qb8
>
>{Equalizing}
>
>23. Qe3 Qd6 24. Nh4 Be6 25. Rh1 Rd8 26. Ng6+ Kg7 27. Nf4 Bf5 28. Nce2 Ne7 29.
>Ng3 Kh8 30. Nxf5 Nxf5 31. Qe4 Qd7
>
>{We are showing draw scores for several moves now, and are now pondering 32.
>Qg2, which for a while looks like flat draw, until we see 32... Qxd4 breaking
>out. This should be interesting to analyze. Of course there's the more direct
>32. Ng6+. Black's next move, however, is no way to draw. I don't think Kasparov
>was conceding a draw yet, either}
>
>32. Rh5 Nxd4
>
>{Kasparov had 10 minutes to time control here, and was behing the curtain of his
>private corner. When I played 32... Nxd4, he was audibly dismayed while still
>behind his curtain without coming to the board! He then spent a full 5 minutes
>reprimanding himself, and with 5 minutes left, got back to thinking about the
>position}
>
>33. Ng6+
>
>{Deep Junior suggests 33.Rh1 and it's still playable for white. Does it work ? I
>don't know, but I think Kasparov didn't consider it for even a second. It seems
>that human players are incapable of accepting the price of a mistake and
>rethinking the position objectively. The knight manouevre gets it over quickly}
I think that the problem was the time trouble of kasparov.
I suspect that with more time he could find Rh1.
I have 2 ideas about changing the time control for future events to help humans:
1)I believe that Fisher time control is better for humans than 2 hours/40 moves
because humans often have the problem of not being able to divide their time
correctly.
I think that 1 hour per game+2 minutes per move may give better chances for
humans.
2)Another idea that I thought about is to change the rules of the time control
so
the player does not lose on time immediately in case of using too much time.
It is possible to decide that for every minute that you think more than you are
allowed you are losing 1% of your points so if you think 100 minutes more than
the time that you are allowed you are losing on time.
A player who get a winning position but blunders in time trouble may feel that
it is not fair.
My idea may allow the player to get most of the point.
If he is using only 10 minutes more than the time that he is allowed he will get
90%.
Uri
This page took 0 seconds to execute
Last modified: Thu, 15 Apr 21 08:11:13 -0700
Current Computer Chess Club Forums at Talkchess. This site by Sean Mintz.