Author: Robert Hyatt
Date: 15:56:35 11/25/00
Go up one level in this thread
On November 25, 2000 at 16:30:02, kurt wrote: >On November 25, 2000 at 14:26:09, Uri Blass wrote: > >>On November 25, 2000 at 13:27:32, Robert Hyatt wrote: >> >>>On November 25, 2000 at 11:01:51, Uri Blass wrote: >>> >>>>On November 25, 2000 at 10:13:40, Thorsten Czub wrote: >>>><snipped> >>>>>Nf6 is a typical bluff move. it pushes the problem over the horizont. >>>>>a horizont problem. this time a positive. >>>>>+ thats the way the new paradigm programs win. >>>>> >>>>>they see a win or a good move. Nf6. >>>>>they play it. >>>>>like a human beeing directed by ideas and illusions. >>>> >>>>I doubt how many humans will play Nf6 in a game. >>>>I believe that most of the strong players will not do it. >>>>They know that they can have a positional advantage with no risk and they will >>>>be afraid to sacrifice a piece if they do not see at least a forced draw. >>>> >>>>I believe that most humans are going to play Ng3. >>>> >>>> >>>>>creative thinking. there is never really an accurate way to win life ! >>>>>but by doing something and having the initiative you often increase >>>>>the chances to win. but - it is risky. if GT would have played >>>>>Nf6 against (say) GOD Nf6 would only have been a nearly forced draw. >>>> >>>>If white can win by another move instead of Nf6+ then Nf6+ is a mistake and when >>>>the opponents will get better playing this move is not going to lead to a win >>>>against them. >>>> >>>>>but the bean counters on the other side of the board are not god. >>>>>they have horizont problems too and this is the reason GT wins although >>>>>it plays smashing inaccurate sacs. >>>> >>>>I believe that the evaluation after search should be accurate. >>>> >>>>My opinion is that not accurate static evaluation can be a good idea only if it >>>>helps to get more accurate evaluation after search. >>>> >>>>When I play correspondence games I expect my opponents to be accurate and if I >>>>find that Nf6+ is leading to a draw then I play another move that gives me a >>>>better chances. >>>>> >>>>>whatever. the games are impressing IMO. >>>>> >>>>>If you have a program that plays accurate, it would e.g. not have played >>>>>Nf6 and other moves, and it would maybe not risk anything. >>>>>it would not risk something because it has computed that this risk is not >>>>>working. >>>>>you get a genius-program. plays boring , but accurate. never doing anything. >>>>>waiting for a mistake of the opponent. >>>> >>>>I disagree. >>>> >>>>If there are 2 moves that are leading to a draw an accurate program can choose >>>>one of them that is a sacrifice. >>>>> >>>>>this is one reason i do believe bob hyatt is wrong. he believes if crafty >>>>>is accurate it would play better chess. i don't think so. >>>> >>>>I think that no program is accurate and that no program is going to be accurate. >>>> >>>>If crafty is going to play accurate then it is never going to lose in chess. >>>>The fact that it is losing is a proof that crafty has no accurate evaluation. >>>> >>>> >>>>>crafty would not do anything. like a human beeing sitting in his chair, >>>>>completely >>>>>autistic because he had considered anything in forward and have found out that >>>>>life >>>>>is dangerous and therefore better not move ONE step forward-. >>>>>cause driving in the car is dangerous. >>>>>better NOT drive. and eating is dangerous. could be poisened. and sleeping is >>>>>dangerous, because you have eyes closed. everything is dangerous. so better >>>>>doing nothing. >>>>>and thats what crafty is mainly doing. accurate doing nothing. >>>>>if crafty would be a human beeing, you would call him ill. >>>> >>>>I disagree. >>>> >>>>There are humans who never sacrifice in chess and I do not call them ill. >>>> >>>>Crafty is not accurate and there are a lot of cases when it evaluates wrong and >>>>this is one of the reasons that Crafty lose games(it can sacrifice the king >>>>safety for a pawn because of wrong evaluation function that say that the pawn >>>>has bigger value). >>>>Gambittiger is also not accurate and can do the opposite mistake. >>>> >>>>The fact that Gambittiger is better than Crafty is not a proof that Gambit is >>>>more accurate about it because gambit is better in tactics and also better in >>>>the endgame. >>>> >>>>Uri >>> >>> >>>I think GT is definitely better in tactics. I don't think it is better in >>>endgames. It has several important pieces of information totally missing. >>>I have posted a game or two showing serious ones. >> >>My impression was based on some of thorsten games that he posted here. >> >>Here are two games when Crafty lost in the endgame when the evaluation of it was >>close to equal during the endgame. >> >>These games are 2 of the 12 games that were posted by thorsten some weeks ago. >> >>Crafty lost at least one more game in the endgame but i prefered to post only >>these 2 games because it is clear that crafty's evaluation was of something that >>is close to 0 at move 40. >> >>The problem in the second game seems to be null move problem and Crafty was in >>zunzwang at move 77. >> >> >> >>[Event "60/60"] >>[Site "k6-400"] >>[Date "29/10/2000"] >>[Round "1-1"] >>[White "Gambit-Tiger1.0"] >>[Black "Crafty17.13"] >>[Result "1-0"] >> >>1. e4 c5 2. Nf3 d6 3. d4 cxd4 4. Nxd4 Nf6 5. Nc3 a6 6. Bg5 >>e6 7. f4 Qb6 8. Qd2 Qxb2 9. Rb1 Qa3 10. f5 Nc6 11. fxe6 >>fxe6 12. Nxc6 bxc6 13. Be2 Be7 14. O-O O-O 15. Rb3 Qc5+ >>16. Be3 Qe5 17. Bf4 Nxe4 18. Nxe4 Qxe4 19. Bxd6 Rxf1+ >>20. Kxf1 Bxd6 21. Qxd6 Qd5 22. Rd3 Qxd6 23. Rxd6 c5 24. c4 >>Rb8 25. Kf2 Kf7 26. Rc6 Ke7 27. Rxc5 Kd7 28. Ra5 Rb2 >>29. Ke3 Bb7 30. a4 Rb3+ 31. Kd4 Ke7 32. Bd1 Rb2 33. Bf3 >>Bxf3 34. gxf3 Rxh2 35. Rxa6 Rh4+ 36. Kc5 Rf4 37. a5 Rxf3 >>38. Rc6 Ra3 39. Kb6 e5 40. Rc7+ Kf6 41. a6 e4 42. a7 e3 >>43. Rc8 e2 44. Re8 Rb3+ 45. Kc6 Ra3 46. Kb7 Rb3+ 47. Ka8 >>Rb2 48. c5 Rd2 49. Re3 Kf5 50. c6 Kf4 51. Rxe2 Rxe2 52. c7 >>Rc2 53. Kb7 Rb2+ 54. Kc6 Rc2+ 55. Kd6 Rd2+ 56. Ke7 Ra2 1-0 >> >> >>[Event "60/60"] >>[Site "k6-400"] >>[Date "29/10/2000"] >>[Round "3-1"] >>[White "Gambit-Tiger1.0"] >>[Black "Crafty17.13"] >>[Result "1-0"] >> >>1. d4 d5 2. c4 c6 3. Nf3 Nf6 4. Nc3 dxc4 5. a4 Bf5 6. e3 e6 >>7. Bxc4 Bb4 8. O-O Nbd7 9. Nh4 O-O 10. f3 Bg6 11. Nxg6 hxg6 >>12. Qc2 Qe7 13. e4 c5 14. Be3 cxd4 15. Bxd4 Bc5 16. Qf2 >>Bxd4 17. Qxd4 Rfd8 18. Bb5 a6 19. Be2 e5 20. Qe3 Qc5 >>21. Qxc5 Nxc5 22. Bc4 Rd4 23. b3 Nd3 24. Ne2 Rd7 25. Rfd1 >>Rad8 26. Nc3 Kf8 27. Rab1 g5 28. h3 g6 29. Rd2 Nf4 30. Rxd7 >>Rxd7 31. Rb2 Ke7 32. Kf1 Ne8 33. Rc2 Nc7 34. Kf2 Nce6 >>35. Ke3 Nd4 36. Rb2 Kd6 37. a5 f6 38. Na4 Rc7 39. Nb6 Rc6 >>40. b4 Ke7 41. Rd2 Kf8 42. Kf2 Kg7 43. Kg3 Rc7 44. Kh2 Kf8 >>45. Rf2 Rh7 46. Bf1 Rc7 47. g3 Nh5 48. Nd5 Rc6 49. Kg2 Kf7 >>50. Rb2 Kg7 51. b5 axb5 52. Bxb5 Rc5 53. Bd3 Rxa5 54. Rxb7+ >>Kh8 55. Bc4 Rc5 56. Rb4 f5 57. Kf2 Nc6 58. Ra4 Kg7 59. Bd3 >>Nf6 60. Nc7 Rc3 61. Ne6+ Kh6 62. Ke2 f4 63. Kd2 Rb3 64. g4 >>Rb8 65. Kc3 Nd7 66. Ra6 Nd8 67. Nxd8 Rxd8 68. Ra5 Rc8+ >>69. Kb4 Rb8+ 70. Rb5 Re8 71. Be2 Re6 72. Rd5 Nf6 73. Ra5 >>Re8 74. Ra6 Nd7 75. Bb5 Rb8 76. Re6 Rb7 77. Re8 Kg7 >>78. Re7+ 1-0 >> >>Uri > >Yes you are right. Crafty lacks instructions from middle to end game face. >Last game above shows clearly that no instruction are given to deminish >the bishop strength by placing pawn on squares of opposite color. This is incorrect. Crafty _does_ have an eval term for this, if you look in evaluate.c... It has been there for a couple of years, at least. >And,how >to maximize its own strength. Like move 27...g5 /g6 etc. and does not find >b4 for its night and force exchange instead of long king march into nowwhere. >I hope Bob reads this concern of mine and provide instructions so that >crafty can cope with such simple positions. >Thanks for your effort to highlight crafts weakness. >kurt.
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.