Author: Robert Hyatt
Date: 14:59:16 11/17/03
Go up one level in this thread
On November 17, 2003 at 16:21:32, Vincent Lejeune wrote:
>On November 17, 2003 at 13:01:14, Robert Hyatt wrote:
>
>>On November 17, 2003 at 10:49:25, martin fierz wrote:
>>
>>>On November 17, 2003 at 10:27:29, Robert Hyatt wrote:
>>>
>>>>On November 17, 2003 at 09:31:28, martin fierz wrote:
>>>>
>>>>>On November 17, 2003 at 09:19:18, Robert Hyatt wrote:
>>>>>
>>>>>>On November 17, 2003 at 07:34:51, Albert Silver wrote:
>>>>>>
>>>>>>>>That's not a place where _I_ would have resigned, for the record...
>>>>>>>
>>>>>>>
>>>>>>>I can't attest to it, as I don't receive it in Brazil, but comments in the ICC
>>>>>>>chat said that at the end the ESPN coverage had been cut. Combined with the fact
>>>>>>>that I also don't believe GK was going to miss the comp, I'd say that's a pretty
>>>>>>>reasonable conspiracy theory... :-)
>>>>>>>
>>>>>>> Albert
>>>>>>
>>>>>>
>>>>>>I wasn't watching ESPN's coverage. I was online on ICC with Crafty running
>>>>>>in channel 211 as usual. The resignation came with crafty at +2, which seems
>>>>>>_way_ early, unless there is some sort of agreement that when Fritz goes
>>>>>>down -2 it _must_ resign...
>>>>>
>>>>>you seem to believe that the number the eval spits out has something to do with
>>>>>how hard it is to win a position... that isn't true ;-)
>>>>>
>>>>>cheers
>>>>> martin
>>>>
>>>>
>>>>I seem to remember "meat makes mistakes". I might resign at -5 or something,
>>>>but not at -2. Remember that Kasparov had good positions in games 1 and 2.
>>>>In game two he lost, rather than Fritz "winning". I'd want to have him
>>>>demonstrate that he would not do the same thing in game 3. :)
>>>
>>>meat makes mistakes *much* more frequently when either
>>>- meat is in time trouble
>>>- the position is complicated
>>>both was not the case here. it was the appropriate moment to resign. i guess if
>>>you play out a few more moves on your machine, as suggested by the commentators
>>>(or basically, exchange a few more pieces by invading on the a-file), crafty
>>>will soon say +3 or more for white.
>>>
>>>cheers
>>> martin
>>
>>
>>My problem with resigning so early is that it is not that unlikely that
>>white makes a small mistake. Then another.
>>
>>I remember Levy playing chess 4.x in the late 70's, and it found a crushing
>>sacrifice against him. He said "The position is totally hopeless for me here"
>>in his later analysis. Then a few moves later he said "it is not not totally
>>hopeless, just hopeless." And eventually it became a draw if I recall
>>correctly.
>>
>>A score of +2 is not winning, IMHO, even though in this position white has all
>>the play. However, the game is wearing on well beyond 40 moves, white is
>>tired. Black is not. The game pace is speeding up. Etc...
>>
>>I don't see any good reason for not playing on a few more moves until it
>>is "completely hopeless" not just "sad". :)
>
>here's the game from Levy's bet
>
>[Event "The Levy Bet 1"]
>[Site "Reading"]
>[Date "1978.08.??"]
>[Round "1"]
>[White "Levy, D.N.L.."]
>[Black "CHESS 4.7"]
>[Result "1/2-1/2"]
>[ECO "A07"]
>[Annotator "DNL"]
>[PlyCount "127"]
>[EventDate "2003.01.10"]
>
>{Notes from Levy & Newborn (1991), "How Computers Play Chess". W.H.Freeman &
>Co. ISBN 0-7167-8121-2, with additions by CA == Computer Analyst} 1. g3 d5 2.
>Bg2 e5 3. d3 Nf6 4. Nf3 Nc6 5. O-O Bd7 6. b3 Bc5 7. Bb2 Qe7 8. a3 e4 9. Ne1 O-O
>10. d4 Bd6 11. e3 {When faced with a strong computer program, I try to play
>the opponent, not the position. It was my plan to create a situation in which
>nothing was happening, and then to expand gradually on the queenside.
>Unfortunately, the program had learned how to attack on the kingside.} 11...
>Ng4 $1 12. h3 $4 (12. c4 {must be played, but even then} 12... Qg5 {
>is strong. I had completely overlooked Black's 13th move.}) 12... Nxe3 $3 {
>The program replied instantly, indicating that it had expected 12.h3 and had
>already worked out its replywhileIwasthinking!} 13. fxe3 Qg5 14. g4 {Realising
>that I was completely busted, I thought my only hope was tosacrifice the
>exchange in order to trade off the Queens.} 14... Qxe3+ 15. Rf2 Bg3 16. Qe2
>Qxf2+ {Of course, it would be crushing to take with the Bishop and keep the
>Queens on so that my King would die of exposure, but the program knows that it
>should trade down when materially ahead.} 17. Qxf2 Bxf2+ 18. Kxf2 f5 $1 19.
>gxf5 Ne7 20. c4 Rxf5+ 21. Kg1 c6 22. Nc3 Rh5 23. Kh2 Rf8 24. Nd1 Ng6 25. Rc1
>Bxh3 $1 {I had seen this coming but was powerless to prevent it.} 26. Bxh3 Rf1
>27. Ng2 Rf3 28. cxd5 Rhxh3+ 29. Kg1 cxd5 30. Rc8+ Nf8 $2 31. Bc3 Rd3 32. Nde3
>Rhxe3 33. Nxe3 Rxe3 34. Bb4 {My first threat of the game, but I was not too
>happy about being three pawns down.} 34... Rf3 35. Rd8 h6 ({
>The program no doubt saw that} 35... Rf5 36. Bxf8 Rxf8 37. Rxd5 {would be follo
>wed by the win of the e-Pawn or the b-Pawn and probably did not analyse past
>this point.}) 36. Rxd5 Rxb3 37. Rd8 Rf3 38. Ra8 g5 39. d5 h5 40. d6 Kg7 41.
>Rxa7 Rf7 42. Ra5 {Suddenly the position is no longer totally hopeless - it is
>merely rather hopeless.} 42... Kf6 43. Bc3+ Kg6 44. Re5 Rf3 45. Bb4 Rf4 46. Re7
>Rf7 47. Rxe4 Rd7 48. Re7 h4 49. Kg2 g4 50. Kh2 $6 (50. Bc5 {
>would have prevented} 50... b6) 50... b6 51. Kg2 Rd8 {
>Help! Black is beginning to untangle its pieces.} 52. a4 Nd7 53. a5 Nf6 ({
>I was expecting} 53... bxa5 54. Bxa5 Ra8 {, and hoping that} 55. Bc3 {
>might hold.}) 54. axb6 $1 {Material equality at last. Now I thought I had a
>draw, and if my opponent made one more mistake, then maybe ...} 54... Nd5 55.
>b7 $1 {At this point in the proceedings the compute got sick and the doctors
>were called. Twenty-five minutes later, with the program still having plenty
>of time on its clock, it played} 55... Nxe7 $1 {A brilliant decision. CHESS 4.
>7 had probably found the only way for Black to draw!} (55... Nxb4 {
>does nothing for Black, if only because of something like} 56. Re4 Nc6 57.
>Rxg4+ Kh5 58. Rc4) ({also} 55... Nf4+ 56. Kh1 {looks good for White.}) 56. dxe7
>Rh8 $1 ({On} 56... Re8 57. Ba5 {wins}) 57. Bd6 ({Fritz6, 2003} 57. Bc3 h3+ 58.
>Kh1 Rb8 59. Be5 Re8 60. Bc7 Rxe7 61. b8=Q Re1+ 62. Kh2 Re2+ 63. Kg1 Re1+ 64.
>Kf2 Re6 65. Qg8+ Kf5 66. Qf7+ Rf6 67. Qh5+ (67. Qd5+) 67... Ke6+ 68. Kg3 Kd7
>69. Qxg4+ Kxc7 70. Qg7+ Kc8) 57... Kf6 58. b8=Q Rxb8 59. Bxb8 {
>and the EGTs say 'draw'} 59... Kxe7 60. Bf4 Kf6 61. Bd2 Kg6 62. Be1 Kg5 63. Bf2
>Kh5 64. Be1 {And David Slate offered me a draw on behalf of his program. This
>was a remarkable game, and the first time a computer program had ever drawn
>with an International Master under tournament conditions. Before the match
>David Slate had some doubts as to whether his program was ready to play me,
>but this game removed them.} 1/2-1/2
>
>
>All the games are here :
>http://www.cs.unimaas.nl/icga/news/events/pastevents/GK-DJ%20event/DeepBriefing.html
That was what I remembered. :)
What is amazing is to watch a GM get into wild tactics vs a computer, and
win a piece. Then a few moves later, he loses a pawn. But he is still
winning. Then another pawn. And before you know it...
That's why I thought resigning at +2 was a bit premature, to put it kindly.
This page took 0.01 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.