|
Main
Date: 28 Jan 2008 10:22:13
From: pialogue
Subject: Solving Chess
|
I realize that not everyone is interested in the concept of "solving chess" due to its many complexities and variations of "midgame" play. However, what if the SECRET to solving chess is in the "draw" dynamic? In other words what if the solution to chess was in playing perfect "defense"? Would not a perfectly played game of chess NOT result in a draw? How else could it end? The ONLY way that someone "wins" is if a "mistake" was made, right? Therefore, if someone wins then that was NOT a "perfect" game played by both sides. It has been said that it is possible for white to force a draw in a number of ways but I am talking about a "perfect game" here where both sides apply perfect strategy which is oriented or strategized toward long-term defense more than offense. I know that some people think that the game would then just be "stuck" with no one able to move but this dynamic plays into what I call "viable strategy" where a stuck game draw is the same as a loss for BOTH sides. In other words, a "true" draw under the "perfect" scenario ends with ONLY the two Kings on the board. So, what are your thoughts on "Solving Chess"? Any thoughts on the best chess game ever played by either human or computer would also be greatly appreciated. Thanks for your potential participation in this intellectual inquiry.
|
|
|
Date: 11 Feb 2008 10:42:49
From: jefk
Subject: Re: Solving Chess??? (it's a draw!)
|
On Feb 11, 7:27 am, jefk <[email protected] > wrote: > the engine still is more important, but i expect > this effect will later converge to almost 100 % draws. as it's a bit vague, let's illustrate this further; what i wrote that Rybka probably draws against Hydra is just an assumption, from estimating playing strength; now i've seen from their site that Donninger is busy with a newer version, and this new Hydra, lets call it Hydra-plus, might beat Rybka again, also with black, as engines still are more important than book lines but this is not the point; the question is how would a perfect engine play with white vs black, and then again i expect that hydra would draw against itself, using of course high quality/solid black book lines. but even the new 'Hydraplus' wont be perfect yet, lets assume we are near (almost) perfect chess, lets say a ''Superhydra'' would play (999.99999999999 %) 'perfect' chess, then i indeed would expect black to draw when this engine would play against itself; book lines would have become more important, but i dont expect very much, i.e no siginificant, but only tiny more improvements in the main book lines which i've already found out, often up to move 25 or so also, when playing similar almost perfect engines against each other, the draw percentage theoretically would approach 100 %, that is, if neither side is resorting to 'outbooking' tricks, ie assuming almost perfect opening books for both sides. reason, well maybe indeed as Sanny pointed out, such games might get much longer, and then the 50 move draw rule will more often play a role, and certainly (apparently) when approaching the 4500 move game length jef PS1 now posted only to chess.analysis (and for mathematical logical discussions i still think that eg. the chess thinkers forum on icdchess would be more appropriate, as other wise we would get off topic) PS2 correction to what i wrote earlier > > with a ryba on a fast comp, > playing against the -even stronger- supercomp/engine > Hydra, i would still expect a lot of draws. wasnt entirely correct i must admit, either the old Hydra is stronger then rybka, then it still might win games with black, or its equally strong then indeed we might expect draws
|
|
Date: 10 Feb 2008 22:33:35
From: jefk
Subject: Re: Solving Chess??? (it's a draw!)
|
On Feb 10, 7:44 pm, Sanny <[email protected] > wrote: > Are you still facing any problem in playing? > no i managed to get in, although i got an email asking me to verify the verification code, and when i did that i got the message 'verification code and user id dont match''; get i could simply login, thereafter, so it *does* match.. >Ask you spam filter to allow mails from that email. i've simply reset the settings eg accepting the sender for that particular email >Can you send me the details why the Spam > Filter Stopped the mail? What was the reasion for filtering it. dont know, its a filter at the internet provider, and these filters are never perfect; it probably recognizes you getclub.com is a commercial site > > What is your webserver name and which filter do you use? well you can see which provider people have from their email, so you have to ask such providers which kind of filters they use :) jef PS i found the applause after each chess move annoying, and again, i hope you can implement this 3 move drawing rule. i also suspect the beginner level now has higher rating in fact than the 1000 you have estimated, but maybe that's based on standard time controls when a human would play eg a long game of 90 minutes; most humans wont have the patience to play such game, and would only take about 10 minutes, eg. blitz to play a game of chess at your site > > Bye > Sanny > > Play Chess at:http://www.GetClub.com/Chess.html
|
|
Date: 10 Feb 2008 22:27:13
From: jefk
Subject: Re: Solving Chess??? (it's a draw!)
|
On Feb 10, 7:46 pm, [email protected] wrote: > > You are proving that players with equal ratings draw each other. > What's the surprise that a computer, playing itself with the same > evaluation functions, draws? the point was not the engine, but the opening lines, played with the strongest engine possible. there's simply no other engine as strong as Rybka. if i would play inferior opening lines, played with the same engine, you still can expect to see wins for white (in case of inferior line for black, eg an unsound gambit like the Latvian) or wins for black (in case of inferior lines for white, eg an unsound gambit, eg e4 g5?) i've also played rybka vs other engines ofcourse, but then the effect of opening theory is ginal, the engine still is more important, but i expect this effect will later converge to almost 100 % draws. as i wrote earlier, with a ryba on a fast comp, playing against the -even stronger- supercomp/engine Hydra, i would still expect a lot of draws. The higher the average rating between the two engines, the higher the drawing gin, provided the opening lines are correct. so in perfect chess, all games will be draw.
|
|
Date: 10 Feb 2008 10:46:05
From:
Subject: Re: Solving Chess??? (it's a draw!)
|
On Feb 10, 9:35=A0am, jefk <[email protected] > wrote: > below some drawn endgames with standard time control > (5 man egtb), dual core AMD4200x64 > (in first two games you see 50 move rule, > thereafter insufficient material to mate), > some clear examples of the drawish nature of chess > (black didnt play 100 % perfect yet, lets say 99.8 % perfect, > white got a small advantage at least in the Rybka score, > but obviously couldnt win in the endgames(s) : > > [Event "Test_long games"] > [Date "2008.02.09"] > [Round "24"] > [White "Rybkav2.3.2a.x64 [mainbk [001]"] > [Black "Rybkav2.3.2a.x64 [mainbk"] > [Result "1/2-1/2"] > [ECO "C94"] > [WhiteElo "3200"] > [BlackElo "3200"] > [PlyCount "271"] > [TimeControl "5400"] > > 1. e4 e5 2. Nf3 Nc6 3. Bb5 a6 4. Ba4 Nf6 5. O-O Be7 6. Re1 b5 7. Bb3 > d6 8. c3 > O-O 9. h3 Nb8 10. d3 Nbd7 11. Nbd2 Bb7 12. Nf1 Nc5 13. Bc2 Re8 14. Be3 > Bf8 15. > Ng3 g6 16. b4 Ne6 17. a4 Bg7 18. Bb3 Qe7 19. Qd2 Qd7 20. Qc2 Rec8 21. > d4 exd4 > 22. cxd4 d5 23. e5 Ne4 24. Nxe4 dxe4 25. Ng5 Nxg5 26. Bxg5 Bf8 27. > Rxe4 Qf5 28. > Rg4 Qxc2 29. Bxc2 Bxb4 30. axb5 axb5 31. Rb1 h5 32. Rg3 c5 33. e6 Ra6 > 34. Bf4 > Kf8 35. exf7 Kxf7 36. Rg5 Ba8 37. Rd1 c4 38. Rxb5 Bd6 39. Bxd6 Rxd6 > 40. d5 c3 > 41. f4 Kf6 42. Kf2 Rcd8 43. Be4 c2 44. Bxc2 Bxd5 45. g4 hxg4 46. hxg4 > g5 47. f5 > Bc4 48. Rc5 Rd2+ 49. Kg3 Rxd1 50. Bxd1 Rd3+ 51. Kf2 Bf7 52. Be2 Ra3 > 53. Bf3 > Ra2+ 54. Kg3 Ra3 55. Kg2 Ra2+ 56. Kh3 Rd2 57. Rc7 Rd3 58. Kg2 Ra3 59. > Kf2 Ra4 > 60. Kg3 Ra3 61. Rd7 Bc4 62. Kf2 Rd3 63. Rc7 Bf7 64. Rc1 Ra3 65. Rd1 > Ra2+ 66. > Kg3 Ra3 67. Rd4 Ke5 68. Rd8 Rb3 69. Rd1 Bc4 70. Re1+ Kf6 71. Kf2 Rb2+ > 72. Kg1 > Rb3 73. Rc1 Bd3 74. Rc8 Rb4 75. Kf2 Rf4 76. Rc6+ Ke5 77. Rc5+ Kf6 78. > Ke3 Bxf5 > 79. gxf5 Rxf5 80. Rc6+ Kg7 81. Be4 Rf6 82. Rc7+ Rf7 83. Rc5 Kf6 84. > Rc6+ Ke5 > 85. Rg6 Ra7 86. Rxg5+ Kd6 87. Rd5+ Ke6 88. Rf5 Ra3+ 89. Kf4 Ra7 90. > Rc5 Rd7 91. > Ke3 Kd6 92. Ra5 Rd8 93. Bd5 Ke5 94. Bc4+ Kf6 95. Ke4 Rf8 96. Rc5 Kg6 > 97. Bd3 > Rd8 98. Rc6+ Kg5 99. Ke3 Rg8 100. Be4 Rg7 101. Rc5+ Kg4 102. Rf5 Ra7 > 103. Rf6 > Rg7 104. Rf1 Kg5 105. Rf5+ Kg4 106. Rb5 Rf7 107. Rb6 Rg7 108. Rb1 Kg5 > 109. Rb5+ > Kg4 110. Bf3+ Kg3 111. Rb4 Re7+ 112. Be4 Kg4 113. Rb2 Kg5 114. Rg2+ > Kf6 115. > Kf4 Re6 116. Bf5 Rd6 117. Ke4 Rd1 118. Rg6+ Kf7 119. Ra6 Rb1 120. Be6+ > Ke7 121. > Bd5 Rh1 122. Kd4 Rg1 123. Re6+ Kd7 124. Rb6 Ke7 125. Rb7+ Kd6 126. Bf3 > Rg5 127. > Rb6+ Ke7 128. Ra6 Kf7 129. Be4 Kg7 130. Ra1 Rh5 131. Bf5 Rh4+ 132. Kd5 > Kf6 133. > Bg4 Kf7 134. Kc6 Kf6 135. Rf1+ Ke7 136. Rf4 {50 moves rule} 1/2-1/2 > > ' > > [Event "Test_standard"] > [Date "2008.02.10"] > [Round "3"] > [White "Rybkav2.3.2a.x64 [mainbk"] > [Black "Rybkav2.3.2a.x64 [mainbk [001]"] > [Result "1/2-1/2"] > [ECO "D75"] > [WhiteElo "3200"] > [BlackElo "3200"] > [PlyCount "300"] > [TimeControl "5400"] > > 1. d4 Nf6 2. c4 e6 3. Nf3 b6 4. g3 Bb7 5. Bg2 c5 6. Nc3 cxd4 7. Qxd4 > Nc6 8. Qf4 > Bc5!N 9. O-O O-O 10. Be3 Qe7 11. Rad1 Rab8 12. Ng5 Rfd8 13. b3 Bb4 14. > Nb5 d5 15. > cxd5 Nxd5 16. Bxd5 exd5 17. Nf3 Rbc8 18. Rc1 Bc5 19. Rfd1 a6 20. Nbd4 > Nxd4 21. > Bxd4 f6 22. Bxc5 bxc5 23. Rc2 g5 24. Qd2 Qe4 25. Ne1 Kg7 26. Rdc1 c4 > 27. f3 Qe6 > 28. Rc3 Qb6+ 29. Kf1 a5 30. Nc2 a4 31. Nd4 cxb3 32. axb3 a3 33. Ra1 > Rxc3 34. > Qxc3 Ra8 35. Ra2 h5 36. Qe3 Kf7 37. Qd3 Kf8 38. Nc2 Bc8 39. Qd4 Qc6 > 40. Kf2 Kg7 > 41. Qe3 g4 42. Qe7+ Kg6 43. Nd4 Qd7 44. Qc5 Qa7 45. Qxa7 Rxa7 46. Nb5 > Ra5 47. > Nxa3 gxf3 48. exf3 d4 49. Ra1 Rc5 50. Nc4 Be6 51. Ra4 Rb5 52. Nd2 d3 > 53. Ke3 > Bxb3 54. Ra6 Bd5 55. Kxd3 Bb7 56. Rd6 Kf5 57. h4 Bd5 58. Ra6 Rb4 59. > Ra5 Ke6 > 60. Ra7 Kf5 61. Ke3 Kg6 62. Rd7 Rb5 63. Ne4 Rb3+ 64. Kf4 Rd3 65. Ra7 > Rd4 66. > Ke3 Rc4 67. Ra1 Bxe4 68. fxe4 Rb4 69. Rc1 Rb3+ 70. Kf4 Rb2 71. Rd1 > Rf2+ 72. Ke3 > Rb2 73. Kf3 Rb3+ 74. Kf4 Rb4 75. Ra1 Rb2 76. Rf1 Rb4 77. Rd1 Rb2 78. > Ke3 Rb3+ > 79. Rd3 Rb2 80. Rd8 Rb3+ 81. Kf4 Rb1 82. Rg8+ Kf7 83. Ra8 Rf1+ 84. Ke3 > Re1+ 85. > Kf3 Rb1 86. Ra7+ Kg6 87. Ra6 Rb3+ 88. Kf4 Rb1 89. Rd6 Kf7 90. Ke3 Re1+ > 91. Kf3 > Rf1+ 92. Kg2 Re1 93. Rd7+ Kg6 94. Kf3 Rf1+ 95. Ke2 Rb1 96. Rd5 Rb3 97. > Rd3 Rb1 > 98. Rd7 Rb3 99. Rd8 Kf7 100. Rd3 Rb5 101. Kf3 Kg6 102. Rd5 Rb3+ 103. > Kf4 Rb1 > 104. Ra5 Rf1+ 105. Ke3 Rg1 106. Kf2 Rb1 107. Ra8 Rb3 108. Rh8 Ra3 109. > Rb8 Kf7 > 110. Rb7+ Kg6 111. Rb5 Rd3 112. Ra5 Rd4 113. Ke3 Rd1 114. Kf3 Rd3+ > 115. Kf4 Rd1 > 116. e5 Rf1+ 117. Ke3 Re1+ 118. Kd4 Rd1+ 119. Kc4 Rg1 120. Ra6 Kf5 > 121. exf6 > Kg6 122. Ra8 Kf7 123. Ra3 Kxf6 124. Rf3+ Ke5 125. Kd3 Rb1 126. Ke3 > Rb3+ 127. > Ke2 Rb2+ 128. Kf1 Ke4 129. Rf2 Rb5 130. Kg2 Ke5 131. Rc2 Kf6 132. Rc4 > Rb3 133. > Rc6+ Kg7 134. Rc5 Kg6 135. Rg5+ Kh6 136. Re5 Kg6 137. Kf2 Ra3 138. > Rg5+ Kh6 > 139. Rd5 Kg6 140. Rd6+ Kf7 141. Kg2 Rc3 142. Rd5 Kg6 143. Rg5+ Kh6 > 144. Rb5 Kg6 > 145. Rd5 Ra3 146. Rg5+ Kh6 147. Re5 Kg6 148. Re6+ Kf7 149. Rc6 Rb3 > 150. Ra6 Rc3 > {1/2-1/2 Arena Adjudication} 1/2-1/2 > > [Event "Test_long games"] > [Date "2008.02.08"] > [Round "13"] > [White "Rybkav2.3.2a.x64 [mainbk"] > [Black "Rybkav2.3.2a.x64 [mainbk [001]"] > [Result "1/2-1/2"] > [ECO "C92"] > [WhiteElo "3200"] > [BlackElo "3200"] > [PlyCount "153"] > [TimeControl "5400"] > > 1. e4 e5 2. Nf3 Nc6 3. Bb5 a6 4. Ba4 Nf6 5. O-O Be7 6. Re1 b5 7. Bb3 > d6 8. c3 > O-O 9. h3 Bb7 10. d4 Re8 11. Nbd2 Bf8 12. a4 h6 13. d5 Nb8 14. c4 bxc4 > 15. Bxc4 > a5 16. Bb5 c6 17. dxc6 Nxc6 18. Nc4 Re6 19. Bd2 Ne7 20. Qc2 Bxe4 21. > Rxe4 Nxe4 > 22. Bxa5 Rxa5 23. Nxa5 d5 24. Nc4 Ng6 25. Ne3 Bc5 26. Rd1 Bxe3 27. > fxe3 Qb6 28. > Re1 Qa5 29. Qc8+ Kh7 30. Qc1 f5 31. Qb1 Qb4 32. Rd1 Qd6 33. Qc2 Qb6 > 34. Re1 Nd6 > 35. b4 Re7 36. Qc6 Rb7 37. a5 Qxb5 38. Qxd6 Qxb4 39. Qxb4 Rxb4 40. Ra1 > Rb7 41. > h4 h5 42. Rc1 Ra7 43. Rc5 e4 44. Ng5+ Kh6 45. g3 Ne7 46. Rb5 Ng8 47. > Rxd5 g6 > 48. Rc5 Nf6 49. Kf1 Nd7 50. Rd5 Kg7 51. Ne6+ Kf7 52. Nd8+ Ke8 53. Nc6 > Ra6 54. > Rd6 Kf8 55. Kg2 Nc5 56. Rxg6 Nb3 57. Re6 Nd2 58. Nd4 Rxa5 59. Rf6+ Kg7 > 60. Rxf5 > Rxf5 61. Nxf5+ Kf6 62. Nd4 Nc4 63. Kf2 Nb2 64. Ke2 Nc4 65. Nb5 Kf5 66. > Kf2 Kg4 > 67. Nc3 Nxe3 68. Kxe3 Kxg3 69. Kxe4 Kxh4 70. Kf4 Kh3 71. Kf3 Kh4 72. > Ne4 Kh3 > 73. Ng5+ Kh4 74. Nh7 Kh3 75. Nf6 Kh4 76. Kf4 Kh3 77. Nxh5 { > Insufficient material} 1/2-1/2 You are proving that players with equal ratings draw each other. What's the surprise that a computer, playing itself with the same evaluation functions, draws?
|
|
Date: 10 Feb 2008 10:44:26
From: Sanny
Subject: Re: Solving Chess??? (it's a draw!)
|
> well i tried to log-in but with difficulties; first > the email with your verification code was blocked > by the spam filter of my server, but with web mail i > managed to find it. Second, after filling in the verification > code i got a new page saying 'page not found'. > anyway, logging in another time solved that problem, > (PS the board appears a to be a bit small, > no not the analysis board, thats very big..) Are you still facing any problem in playing? Ask you spam filter to allow mails from that email. Can you send me the details why the Spam Filter Stopped the mail? What was the reasion for filtering it. What is your webserver name and which filter do you use? Bye Sanny Play Chess at: http://www.GetClub.com/Chess.html
|
|
Date: 10 Feb 2008 06:35:51
From: jefk
Subject: Re: Solving Chess??? (it's a draw!)
|
below some drawn endgames with standard time control (5 man egtb), dual core AMD4200x64 (in first two games you see 50 move rule, thereafter insufficient material to mate), some clear examples of the drawish nature of chess (black didnt play 100 % perfect yet, lets say 99.8 % perfect, white got a small advantage at least in the Rybka score, but obviously couldnt win in the endgames(s) : [Event "Test_long games"] [Date "2008.02.09"] [Round "24"] [White "Rybkav2.3.2a.x64 [mainbk [001]"] [Black "Rybkav2.3.2a.x64 [mainbk"] [Result "1/2-1/2"] [ECO "C94"] [WhiteElo "3200"] [BlackElo "3200"] [PlyCount "271"] [TimeControl "5400"] 1. e4 e5 2. Nf3 Nc6 3. Bb5 a6 4. Ba4 Nf6 5. O-O Be7 6. Re1 b5 7. Bb3 d6 8. c3 O-O 9. h3 Nb8 10. d3 Nbd7 11. Nbd2 Bb7 12. Nf1 Nc5 13. Bc2 Re8 14. Be3 Bf8 15. Ng3 g6 16. b4 Ne6 17. a4 Bg7 18. Bb3 Qe7 19. Qd2 Qd7 20. Qc2 Rec8 21. d4 exd4 22. cxd4 d5 23. e5 Ne4 24. Nxe4 dxe4 25. Ng5 Nxg5 26. Bxg5 Bf8 27. Rxe4 Qf5 28. Rg4 Qxc2 29. Bxc2 Bxb4 30. axb5 axb5 31. Rb1 h5 32. Rg3 c5 33. e6 Ra6 34. Bf4 Kf8 35. exf7 Kxf7 36. Rg5 Ba8 37. Rd1 c4 38. Rxb5 Bd6 39. Bxd6 Rxd6 40. d5 c3 41. f4 Kf6 42. Kf2 Rcd8 43. Be4 c2 44. Bxc2 Bxd5 45. g4 hxg4 46. hxg4 g5 47. f5 Bc4 48. Rc5 Rd2+ 49. Kg3 Rxd1 50. Bxd1 Rd3+ 51. Kf2 Bf7 52. Be2 Ra3 53. Bf3 Ra2+ 54. Kg3 Ra3 55. Kg2 Ra2+ 56. Kh3 Rd2 57. Rc7 Rd3 58. Kg2 Ra3 59. Kf2 Ra4 60. Kg3 Ra3 61. Rd7 Bc4 62. Kf2 Rd3 63. Rc7 Bf7 64. Rc1 Ra3 65. Rd1 Ra2+ 66. Kg3 Ra3 67. Rd4 Ke5 68. Rd8 Rb3 69. Rd1 Bc4 70. Re1+ Kf6 71. Kf2 Rb2+ 72. Kg1 Rb3 73. Rc1 Bd3 74. Rc8 Rb4 75. Kf2 Rf4 76. Rc6+ Ke5 77. Rc5+ Kf6 78. Ke3 Bxf5 79. gxf5 Rxf5 80. Rc6+ Kg7 81. Be4 Rf6 82. Rc7+ Rf7 83. Rc5 Kf6 84. Rc6+ Ke5 85. Rg6 Ra7 86. Rxg5+ Kd6 87. Rd5+ Ke6 88. Rf5 Ra3+ 89. Kf4 Ra7 90. Rc5 Rd7 91. Ke3 Kd6 92. Ra5 Rd8 93. Bd5 Ke5 94. Bc4+ Kf6 95. Ke4 Rf8 96. Rc5 Kg6 97. Bd3 Rd8 98. Rc6+ Kg5 99. Ke3 Rg8 100. Be4 Rg7 101. Rc5+ Kg4 102. Rf5 Ra7 103. Rf6 Rg7 104. Rf1 Kg5 105. Rf5+ Kg4 106. Rb5 Rf7 107. Rb6 Rg7 108. Rb1 Kg5 109. Rb5+ Kg4 110. Bf3+ Kg3 111. Rb4 Re7+ 112. Be4 Kg4 113. Rb2 Kg5 114. Rg2+ Kf6 115. Kf4 Re6 116. Bf5 Rd6 117. Ke4 Rd1 118. Rg6+ Kf7 119. Ra6 Rb1 120. Be6+ Ke7 121. Bd5 Rh1 122. Kd4 Rg1 123. Re6+ Kd7 124. Rb6 Ke7 125. Rb7+ Kd6 126. Bf3 Rg5 127. Rb6+ Ke7 128. Ra6 Kf7 129. Be4 Kg7 130. Ra1 Rh5 131. Bf5 Rh4+ 132. Kd5 Kf6 133. Bg4 Kf7 134. Kc6 Kf6 135. Rf1+ Ke7 136. Rf4 {50 moves rule} 1/2-1/2 ' [Event "Test_standard"] [Date "2008.02.10"] [Round "3"] [White "Rybkav2.3.2a.x64 [mainbk"] [Black "Rybkav2.3.2a.x64 [mainbk [001]"] [Result "1/2-1/2"] [ECO "D75"] [WhiteElo "3200"] [BlackElo "3200"] [PlyCount "300"] [TimeControl "5400"] 1. d4 Nf6 2. c4 e6 3. Nf3 b6 4. g3 Bb7 5. Bg2 c5 6. Nc3 cxd4 7. Qxd4 Nc6 8. Qf4 Bc5!N 9. O-O O-O 10. Be3 Qe7 11. Rad1 Rab8 12. Ng5 Rfd8 13. b3 Bb4 14. Nb5 d5 15. cxd5 Nxd5 16. Bxd5 exd5 17. Nf3 Rbc8 18. Rc1 Bc5 19. Rfd1 a6 20. Nbd4 Nxd4 21. Bxd4 f6 22. Bxc5 bxc5 23. Rc2 g5 24. Qd2 Qe4 25. Ne1 Kg7 26. Rdc1 c4 27. f3 Qe6 28. Rc3 Qb6+ 29. Kf1 a5 30. Nc2 a4 31. Nd4 cxb3 32. axb3 a3 33. Ra1 Rxc3 34. Qxc3 Ra8 35. Ra2 h5 36. Qe3 Kf7 37. Qd3 Kf8 38. Nc2 Bc8 39. Qd4 Qc6 40. Kf2 Kg7 41. Qe3 g4 42. Qe7+ Kg6 43. Nd4 Qd7 44. Qc5 Qa7 45. Qxa7 Rxa7 46. Nb5 Ra5 47. Nxa3 gxf3 48. exf3 d4 49. Ra1 Rc5 50. Nc4 Be6 51. Ra4 Rb5 52. Nd2 d3 53. Ke3 Bxb3 54. Ra6 Bd5 55. Kxd3 Bb7 56. Rd6 Kf5 57. h4 Bd5 58. Ra6 Rb4 59. Ra5 Ke6 60. Ra7 Kf5 61. Ke3 Kg6 62. Rd7 Rb5 63. Ne4 Rb3+ 64. Kf4 Rd3 65. Ra7 Rd4 66. Ke3 Rc4 67. Ra1 Bxe4 68. fxe4 Rb4 69. Rc1 Rb3+ 70. Kf4 Rb2 71. Rd1 Rf2+ 72. Ke3 Rb2 73. Kf3 Rb3+ 74. Kf4 Rb4 75. Ra1 Rb2 76. Rf1 Rb4 77. Rd1 Rb2 78. Ke3 Rb3+ 79. Rd3 Rb2 80. Rd8 Rb3+ 81. Kf4 Rb1 82. Rg8+ Kf7 83. Ra8 Rf1+ 84. Ke3 Re1+ 85. Kf3 Rb1 86. Ra7+ Kg6 87. Ra6 Rb3+ 88. Kf4 Rb1 89. Rd6 Kf7 90. Ke3 Re1+ 91. Kf3 Rf1+ 92. Kg2 Re1 93. Rd7+ Kg6 94. Kf3 Rf1+ 95. Ke2 Rb1 96. Rd5 Rb3 97. Rd3 Rb1 98. Rd7 Rb3 99. Rd8 Kf7 100. Rd3 Rb5 101. Kf3 Kg6 102. Rd5 Rb3+ 103. Kf4 Rb1 104. Ra5 Rf1+ 105. Ke3 Rg1 106. Kf2 Rb1 107. Ra8 Rb3 108. Rh8 Ra3 109. Rb8 Kf7 110. Rb7+ Kg6 111. Rb5 Rd3 112. Ra5 Rd4 113. Ke3 Rd1 114. Kf3 Rd3+ 115. Kf4 Rd1 116. e5 Rf1+ 117. Ke3 Re1+ 118. Kd4 Rd1+ 119. Kc4 Rg1 120. Ra6 Kf5 121. exf6 Kg6 122. Ra8 Kf7 123. Ra3 Kxf6 124. Rf3+ Ke5 125. Kd3 Rb1 126. Ke3 Rb3+ 127. Ke2 Rb2+ 128. Kf1 Ke4 129. Rf2 Rb5 130. Kg2 Ke5 131. Rc2 Kf6 132. Rc4 Rb3 133. Rc6+ Kg7 134. Rc5 Kg6 135. Rg5+ Kh6 136. Re5 Kg6 137. Kf2 Ra3 138. Rg5+ Kh6 139. Rd5 Kg6 140. Rd6+ Kf7 141. Kg2 Rc3 142. Rd5 Kg6 143. Rg5+ Kh6 144. Rb5 Kg6 145. Rd5 Ra3 146. Rg5+ Kh6 147. Re5 Kg6 148. Re6+ Kf7 149. Rc6 Rb3 150. Ra6 Rc3 {1/2-1/2 Arena Adjudication} 1/2-1/2 [Event "Test_long games"] [Date "2008.02.08"] [Round "13"] [White "Rybkav2.3.2a.x64 [mainbk"] [Black "Rybkav2.3.2a.x64 [mainbk [001]"] [Result "1/2-1/2"] [ECO "C92"] [WhiteElo "3200"] [BlackElo "3200"] [PlyCount "153"] [TimeControl "5400"] 1. e4 e5 2. Nf3 Nc6 3. Bb5 a6 4. Ba4 Nf6 5. O-O Be7 6. Re1 b5 7. Bb3 d6 8. c3 O-O 9. h3 Bb7 10. d4 Re8 11. Nbd2 Bf8 12. a4 h6 13. d5 Nb8 14. c4 bxc4 15. Bxc4 a5 16. Bb5 c6 17. dxc6 Nxc6 18. Nc4 Re6 19. Bd2 Ne7 20. Qc2 Bxe4 21. Rxe4 Nxe4 22. Bxa5 Rxa5 23. Nxa5 d5 24. Nc4 Ng6 25. Ne3 Bc5 26. Rd1 Bxe3 27. fxe3 Qb6 28. Re1 Qa5 29. Qc8+ Kh7 30. Qc1 f5 31. Qb1 Qb4 32. Rd1 Qd6 33. Qc2 Qb6 34. Re1 Nd6 35. b4 Re7 36. Qc6 Rb7 37. a5 Qxb5 38. Qxd6 Qxb4 39. Qxb4 Rxb4 40. Ra1 Rb7 41. h4 h5 42. Rc1 Ra7 43. Rc5 e4 44. Ng5+ Kh6 45. g3 Ne7 46. Rb5 Ng8 47. Rxd5 g6 48. Rc5 Nf6 49. Kf1 Nd7 50. Rd5 Kg7 51. Ne6+ Kf7 52. Nd8+ Ke8 53. Nc6 Ra6 54. Rd6 Kf8 55. Kg2 Nc5 56. Rxg6 Nb3 57. Re6 Nd2 58. Nd4 Rxa5 59. Rf6+ Kg7 60. Rxf5 Rxf5 61. Nxf5+ Kf6 62. Nd4 Nc4 63. Kf2 Nb2 64. Ke2 Nc4 65. Nb5 Kf5 66. Kf2 Kg4 67. Nc3 Nxe3 68. Kxe3 Kxg3 69. Kxe4 Kxh4 70. Kf4 Kh3 71. Kf3 Kh4 72. Ne4 Kh3 73. Ng5+ Kh4 74. Nh7 Kh3 75. Nf6 Kh4 76. Kf4 Kh3 77. Nxh5 { Insufficient material} 1/2-1/2
|
|
Date: 10 Feb 2008 06:25:26
From: jefk
Subject: Re: Solving Chess??? (it's a draw!)
|
On Feb 10, 7:47 am, Sanny <[email protected] > wrote: > > I think Master Level at GetClub will play much like perfect game. well there's an important element that's not implemented yet, i've seen that's important the 3 times draw-repitition rule. after you've implemented that you could try to compare the program with commercial programs, and let it play faster at master level (a Russian program Ufim in fact has been programmed to play slower at easy /medium levels, ie wait a few more seconds before moving to simulate longer 'thinking' of the computer opponent > Play Chess at:http://www.GetClub.com/Chess.html well i tried to log-in but with difficulties; first the email with your verification code was blocked by the spam filter of my server, but with web mail i managed to find it. Second, after filling in the verification code i got a new page saying 'page not found'. anyway, logging in another time solved that problem, (PS the board appears a to be a bit small, no not the analysis board, thats very big..) the about chess=draw again: > MATHEMATICAL PROOF BY INDUCTION > So as the Game players strength increases the loss after X moves > reduces. So by Induction as the strength of player playing Chess > increases the loss of piece -> 0 and number of moves after which a > pawn is killed -> INFINITE. > > And we know because of 50 move rule a Chess game cannot exceed 4800 > Moves. So we can conclude a Perfect player will always get Chess to > draw. well i'm not sure whether this 'proof' will be up to rigorous mathematical standards, probably not, as it's not socalled 'complete induction' (Pascal) (if chess is draw for move N=1, assume chess is a draw at move N (eg.59 or 3314..; now can we then 'prove' that it's also a draw at move N+1; if yes, then its proven by complete induction); anyway reasoning along such standards, would be a better way to go i presume than proof by a brute force search (although i've already done something like that for the opening stage up to eg. move 25) But who cares, for example you can prove the ''ideal gas law'' PV=RT by deriving it from the statistical molecule laws by Boltzmann; an interesting mathematical exercise, but you can also derive the law from experiment (although then it appears reality is not behaving completely in line with the mathematics, because of some other effect (non-ideal gas law). In general induction in physics leads to a strong hypothesis, which people can try to falsify; in mathematics its called a conjecture, eg. before Fermats theorem a^N+b^N=c^N (with N integer, only true for N <3) was proven by Wiley it was already known to be true for different values of N, and the conjecture was that is was true. Similar apparently that checkers was known to be a draw as David Kane posted. For chess in mathematical language it's also at least a conjecture that its a draw, which is not placing the burdon of proof on me, but on others imho, ie either to find a rigorous proof in terms of complete induction simultaneous applying also the chess rules (eg 3 times draw rule), or, trying to falsify the meta-abstract empirical finding that chess is a draw by showing winning lines for white against every black defence possible, eg using a quantum computer (well forget it i would say, there are better ways to use computing power) As for your 'induction' using the 50 move draw rule, (didnt now that max length in chess is 4800 moves but this indeed is another indication of its drawish nature), in a next posting i can show two examples of the 50 moves rules, where black didn't even play 100% perfect yet i.e white managed to get a small advantage (which i corrected by introducing novelties for black) but anyway in these games white's advantage already wasnt enough to lead it to a win.. jef
|
|
Date: 09 Feb 2008 22:47:38
From: Sanny
Subject: Re: Solving Chess??? (it's a draw!)
|
> > jefk>to show thatchessis a draw (it is, btw). > > > Liar. > > so if worldchesschampion Anand claimschessis > a draw, he is a liar ? > Surechesscan be a dynamic game sometimes, > whereby the comp evaluation gradually is changing, > but by playing rybka-rybka standard =A0game, not only opening > theory is improved , but the better it gets, the more drawish > the game seems to become. You are correct the Perfect game will be draw. Say White first move. A bad player will play a move Which looses its queen after 10 moves. A little good player will play a move so that his pawn gets killed after 10 moves. A Beginner Level at GetClub will play So that a pwn is lost in 15 moves. Master Level will play a bad move So that a pawn in lost after 30 Moves. MATHEMATICAL PROOF BY INDUCTION So as the Game players strength increases the loss after X moves reduces. So by Induction as the strength of player playing Chess increases the loss of piece - > 0 and number of moves after which a pawn is killed - > INFINITE. And we know because of 50 move rule a Chess game cannot exceed 4800 Moves. So we can conclude a Perfect player will always get Chess to draw. I think Master Level at GetClub will play much like perfect game. Play Chess at: http://www.GetClub.com/Chess.html Bye Sanny Play Chess at: http://www.GetClub.com/Chess.html
|
|
Date: 09 Feb 2008 12:04:10
From: jefk
Subject: Re: Solving Chess??/ a drawish example
|
ok, as i probably will get some reactions again, here's an example with a drawish line, and a but not spectacular little novelty, at move 23; the black 16th move is more important, but i suspect only people who understand a lot of chess are able to see the siginificance of this, and then maybe as well for the drawish nature of chess(*) [Event "Test_long games"] [Site "PC_JEFK"] [Date "2008.02.08"] [Round "10"] [White "Rybkav2.3.2a.x64 [mainbk [001]"] [Black "Rybkav2.3.2a.x64 [mainbk"] [Result "1/2-1/2"] [ECO "C92"] [WhiteElo "3200"] [BlackElo "3200"] [PlyCount "84"] [TimeControl "5400"] 1. e4 e5 2. Nf3 Nc6 3. Bb5 a6 4. Ba4 Nf6 5. O-O Be7 6. Re1 b5 7. Bb3 d6 8. c3 O-O 9. h3 Bb7 10. d4 Re8 11. Nbd2 Bf8 12. a4 h6 13. Bc2 exd4 14. cxd4 Nb4 15. Bb1 c5 16. d5 g6!? { in what usually is considered the mainline of the Zaitsev black plays Nd7?!, after which white is thought to have a slight advantage after 17Ra3!? (but after analyzing 17..g6 -a move often disregarded- this is not the case)} 17. Nf1 Bg7 18. axb5 axb5 19. Rxa8 Qxa8 20. Bf4 {in conventional opening theory for the Zaitsev this was considered an advantage for white, based on two (human) games Gruenfeld(2459) played in 1984, but his opponents then failed to see the response Qd8! better way to protect the pawn on d6 then Qa6?} Qd8! {in Vroom (2200)- Demyak (2205), 1/2-1/2 white played 21 Qd2 , but Ng3 seems better} 21. Ng3 {!N} h5 22. Re3 Nd7 23. Bxd6 Bxb2 24. Rb3 Bg7 25. Bxc5 Nxc5 26. Rxb4 Qa5 27. Qd2 Na6 28. Rb3 Qxd2 29. Nxd2 b4 30. Re3 h4 31. Ngf1 Nc5 32. Nb3 Na4 33. Bd3 Nc3 34. g4 Bh6 35. Rf3 Ba8 36. Nc5 Rc8 37. Nb3 Nxd5 38. exd5 Bxd5 39. Ba6 Ra8 40. Rd3 Bxb3 41. Rxb3 Rxa6 42. Rxb4 {rook+knight with 3 pawns on kingside, vs rook+bishop on kingside, an obvious draw.} f6 { 1/2-1/2 User Adjudication} 1/2-1/2 jef PS of course i have many more examples against 1.e4 e5! And of course i've also looked at eg. Scottish, Italian, etc. For (ie against) 1.d4 i'll might give some example later on the (obvious more mature) forum icdchess.com No need to look at Sicilian or Petrov, as with the Ruy Lopez, especially the Zaitsev,l black seems to be fine. (*) PS2 some may say that a few examples doesnt mean anything, but what i've found is that in the middle game, just because of the 'exponential growth' of the nr of possible playable positions, in case of white winning line, black always can manage to find another option to avoid such a white win. If you don't believe it, you might try it, but its a lot of work, and unless you wouldnt like to re-write the complete Yugoslavian ECO/Encyclopedia of chess openings and wouldnt really recommend it.. (cheers)
|
|
Date: 09 Feb 2008 11:00:47
From: jefk
Subject: Re: Solving Chess??? (it's a draw!)
|
On Feb 7, 11:01 am, NOSPAM <[email protected] > wrote: > > jefk>to show that chess is a draw (it is, btw). > > Liar. so if world chess champion Anand claims chess is a draw, he is a liar ? Sure chess can be a dynamic game sometimes, whereby the comp evaluation gradually is changing, but by playing rybka-rybka standard game, not only opening theory is improved , but the better it gets, the more drawish the game seems to become. who cares about a mathematical proof; can you eg prove the Newtonion law action = minus reaction ? scientific progress consists of paradigm shifts.. By claiming that chess still could be solved in the sense that its a win for white, people are defying centuries of chess theory whereby it's known that long succesful combinations usually only can be made after sufficient positional advantage has been built up. saying that a move as eg. 1 f3 *could* be the winning move simply is ridiculous and that black could also be winning is even more crazy > > Don't let the door hit your ass on the way out, loser. easy to resort to name calling when your posting anonymously, isnt it; haven't they told you sometimes you got bad manners ? (an American i presume; but even worse, you probably cannot play chess at all, judging from your posting) > Then don't make claims of a mathematical/logical > nature that you can't prove. well i dont have to; i'm a physicist, and draw my conclusions from empirical facts. i've played *many* slow comp-comp blitz games on a fast comp, analyzed in detail the resulting book/database, also played and investigated full length standard comp-comp games with novelties, at least for the most drawish book lines.. anyway, i might be back a few months later on the more decent icdchess forum, instead of this usenet, and hopefully with some examples, and/or more arguments to support my empirical 'hunch' that chess is a draw; until then, others (feel free to try) can of course try to falsify this by finding winning lines for white against any possible defence for black.. good luck.. PS according to 'philosopher' P.Feyerabend Galileo Galilei has not really 'proven' that the earth is revolving around the sun (well it does, doesn it; any mathematical solution or hypothesis describing real world eg with Newtonian mechanics should be a simple as possible; you always extend it later with general relativity, but thats makes not much difference ) Also according to this 'philosopher' astrology and witchcraft are just as valid thought systems as mainstream science. well i wonder then how this bloke got his degree..
|
|
Date: 08 Feb 2008 06:37:54
From: Sanny
Subject: Re: Master Level/ Advance Level will play perfect.
|
> >GetClubGame is made just to win against Human Players, I have no > > motive of =A0having it play with other Commercial programs. As it is > > made just for fun and not for competing with other CommercialChess > > Programs. > > =A0 Good. =A0Because you would lose. =A0Badly. > > =A0 -- help bot Today again the GetClub Chess program was improved. So you will find it playing much better than earlier. Bye Sanny Play Chess at: http://www.GetClub.com/Chess.html
|
|
Date: 06 Feb 2008 19:31:43
From: help bot
Subject: Re: Master Level/ Advance Level will play perfect.
|
On Feb 5, 2:24 am, Sanny <[email protected] > wrote: > > Sanny, you already have oodles of information showing that your > > program plays nowhere near perfectly on any of its levels. It is poor- > > to-mediocre at best, atrocious at worst. You are quite clearly NOT > > correct. Time for a reality check. The program is now so strong that in one game, I racked up over ninety moves before finally prevailing! > > So, you are either a liar or a fool. Which is it? Why not both? > I think you have a bit Problem in understanding things. I am neither > liar nor a fool. You are doing unnecesary arguments on a thing that do > not matter much. Ridiculous! What on Earth could possibly be more important than chess strength and chess ratings? (Feeding the hungry? World peace? Puleeeeeeze.) > If you had beaten Master Level then it does not mean it cannot improve > further . It is improving every week. Earlier everyone used to play > with Master level and Win. Now most of them are playing with Beginner > & Easy Level. Why? The time taken by Master has dropped down from > earlier 1 hour to 10 Minutes now. Still people are not willing to play > with Master level as they are afraid of loosing the game. > > Now I find many games where Beginner Level has Won. And I find that > everyones rating is going down. Bingo! Everyone not using a superior program will have a tougher and tougher time of it, and the result will be a narrowing in the gap between GetClub's various levels and the human players. Of course, there will always be a few who RUN! > If the improvement continues as before the day is not far when It will > play as good as Crafty/ Fritz/ Rybka and other Strong Chess Programs. Whoa there, fella! There is a huge difference between besting average human players and competing toe-to-toe with the Kings of chess; some of those programs are impossible to compete against, using your slow Java applet approach. > GetClub Game is made just to win against Human Players, I have no > motive of having it play with other Commercial programs. As it is > made just for fun and not for competing with other Commercial Chess > Programs. Good. Because you would lose. Badly. -- help bot
|
|
Date:
From: Martin Brown
Subject: Re: Solving Chess??? (it's a draw!)
|
|
|
Date:
From: Martin Brown
Subject: Re: Solving Chess??? /perfect play
|
|
| |
Date: 06 Feb 2008 13:31:57
From: David Richerby
Subject: Re: Solving Chess??? /perfect play
|
tin Brown <
|
|
Date: 05 Feb 2008 09:45:22
From: Sanny
Subject: Re: Solving Chess??? (it's a draw!)
|
On Feb 5, 10:41 pm, Guy Macon <http://www.guymacon.com/ > wrote: > Content-Transfer-Encoding: 8Bit > > tin Brown wrote: > >The number of positions that needs to be considered is still far beyond > >anything that a conventional computer can do now or in the future. If > >someone gets a quantum computer to run achessprogram all bets are off, > >but until that day the best we can say is that it might be a draw. > > IBM has a 7-Qbit Quantum Computer that uses Shor's algorithm to > factor the number 15 into 3 and 5. D-Wave Systems claims to have > a working 28-Qbit Quantum Computer, but I have my doubts. Quantum > computers are hard to scale up. There are problems keeping the > Qbits stable long enough to finish the calculation, and as you > add more qubits, the problem gets worse. So we aren't there yet > and may never get there. > > -- > Guy Macon > <http://www.guymacon.com/> How these Quantum Computer work and how we will program them. I suppose it is a Parralel Computer? Bye Sanny Play Chess at: http://www.GetClub.com/Chess.html
|
|
Date: 05 Feb 2008 17:41:04
From: Guy Macon
Subject: Re: Solving Chess??? (it's a draw!)
|
Content-Transfer-Encoding: 8Bit tin Brown wrote: >The number of positions that needs to be considered is still far beyond >anything that a conventional computer can do now or in the future. If >someone gets a quantum computer to run a chess program all bets are off, >but until that day the best we can say is that it might be a draw. IBM has a 7-Qbit Quantum Computer that uses Shor�s algorithm to factor the number 15 into 3 and 5. D-Wave Systems claims to have a working 28-Qbit Quantum Computer, but I have my doubts. Quantum computers are hard to scale up. There are problems keeping the Qbits stable long enough to finish the calculation, and as you add more qubits, the problem gets worse. So we aren't there yet and may never get there. -- Guy Macon <http://www.guymacon.com/ >
|
|
Date:
From: Martin Brown
Subject: Re: Solving Chess??? (it's a draw!)
|
|
| |
Date: 06 Feb 2008 12:44:28
From: jefk
Subject: Re: Solving Chess??? /perfect play
|
tin Brown wrote: > > It is plausible, but it could equally be a first player win. > ok, another thought experiment, also to backup my quasi empirical hunch that the more complicated a symmetrical game, the more likely the outcome of perfect play would be a draw; otherwise there's something wrong with the rules, anyway: lets have a look a soccer; soccer ?? yep, soccer :) does it matter which side has the first kick ? nope, most likely not, it's a long game, 3 hours, and the chances will equalize. Now here's a prediction of the outcome in case of perfect play : 0-0 ! 'ultra weak solution': in case of perfect play, the defence should also be perfect, and -most importantly- also the keeper, in fact, being a very tall and fast guy (or robot ?) he will not let pass through any ball, so there will never be a goal made. Its probably not so difficult to make such a keeper robot from eg. military technology (think about patriot missiles). And then we also have the FIFA off-side rule. Offside rule ? Yep, offside rule, just like the 50 move FIDE endgame rule :-)) (maybe irrelevant, but yet another argument to support a claim) Now you could say such a goalkeeper robot would spoil the game, in case of permanent draws, but thats not the point. THe point is that it would be foolish to think about 22 robo-players, and generate a database -or generate trillions of simulations- to 'prove' that the game would be 0-0, when this already is obvious from empirical facts. In chess, i'm convinced that in coming years we will see more draws in computer-computer games, and gradually the computer chess community will see that chess is a draw. Except those who are member of the flat earth society maybe :) jef
|
| | |
Date: 06 Feb 2008 13:13:16
From: David Richerby
Subject: Re: Solving Chess??? /perfect play
|
jefk <[email protected] > wrote: > ok, another thought experiment, also to backup my quasi empirical > hunch that the more complicated a symmetrical game, the more likely > the outcome of perfect play would be a draw; otherwise there's > something wrong with the rules, anyway: > > lets have a look a soccer; soccer ?? yep, soccer :) 1. Obviously, a perfect defence would never concede a goal. Equally obviously, a perfect attack would never fail to score once it had the ball. This contradiction immediately shows that there's no such thing as `perfect play' in soccer. 2. Your argument didn't involve chess at all but your conclusion did. Dave. -- David Richerby Permanent Moistened Drink (TM): it's www.chiark.greenend.org.uk/~davidr/ like a refreshing juice beverage but it's moist and it'll be there for ever!
|
| | | |
Date: 06 Feb 2008 14:30:48
From: jefk
Subject: Re: Solving Chess??? /rules
|
David Richerby wrote: > 2. Your argument didn't involve chess at all but your conclusion did. > well i mentioned some (speculative) thing which you guys overlooked, namely that *if* the game would *not* be a draw, there must be something wrong with the rules, so the rules then should be changed. Whether its offside, endgame rule, or whatever. I mean, robosoccer (and it exists, at least as table-robo-soccer) wouldnt be fun if the side who starts would always win, isnt it. Let's *assume* (speculative, yes) that the Fide adopted the 50 move endgame rule again to keep chess a draw (although they also can't know that for sure, i agree); then in case that they would be wrong, if eg. a certain David R would find a win for white in 250 moves, then they simply could make more strict rules, eg. making it a 30 move endgame rule, and voila, chess would be a draw again. (but no, i'm not gonna mail Ilmoesjinov :) \ Reduction ad absurdum, yes, by changing the rules as necessary to maintain an equal and honest thus drawish game, so indeed not a perfect solution, but you might see my point. Do we then need to change the rules for fur-in-a-row ? Yep, why not ? But as simple as possible of course. Anyway, glad to see some comon chess sense there that you also suspect chess is a draw. ok, i'll withdraw my statement that that is a 'solution'.. but i still predict we will see more draws after some years, also with freestyle chess;(eg if these guys abandon this bloody Najdorf variation :) best regards, jef
|
| | | | |
Date: 06 Feb 2008 15:48:54
From: David Richerby
Subject: Re: Solving Chess??? /rules
|
jefk <[email protected] > wrote: > David Richerby wrote: >> 2. Your argument didn't involve chess at all but your conclusion did. > > well i mentioned some (speculative) thing which you guys overlooked, > namely that *if* the game would *not* be a draw, there must be > something wrong with the rules, so the rules then should be changed. There are all kinds of games where draws are either very rare or impossible. This in no way means that there's something wrong with the rules of those games. > Whether its offside, endgame rule, or whatever. I mean, robosoccer > (and it exists, at least as table-robo-soccer) wouldnt be fun if the > side who starts would always win, isnt it. But this is completely irrelevant to chess. > Let's *assume* (speculative, yes) that the Fide adopted the 50 move > endgame rule again to keep chess a draw No, let's not assume that. It's a stupid thing to assume. > Reduction ad absurdum, yes No, it isn't reductio ad absurdum. Reductio ad absurdum is proof by contradiction: you assume the opposite of what you're trying to prove and use that to prove a contradiction. For example, I wished to prove that there was no such thing as `perfect soccer'. So I assumed the opposite: that there is perfect soccer. I then noted that a perfect attacker would always score but a perfect defence would always prevent this. This is a contradiction, which shows that the hypothesis that `perfect soccer' exists must be false. Dave. -- David Richerby Disposable Watch (TM): it's like a www.chiark.greenend.org.uk/~davidr/ precision chronometer but you never have to clean it!
|
| | | | |
Date: 06 Feb 2008 14:43:12
From: jefk
Subject: Re: Solving Chess??? /rules
|
jefk wrote: > necessary to maintain an equal and honest thus drawish game, instead of 'honest' i meant *fair* of course, a fair game between to sides should have equal chances for both sides inho. according to GM Adorjan the white statistical advantage is based on psychological factors; i doubt that, as its known that black has less choices, possible moves to make in most -and especially sharp- positions. So this is a problem in any tournament i would say, don't know exactly how the Swiss rules are going, but having seen some comp tourns in Leiden i suspect the black/white side were equally rewarded in case of a win, which is biased; a black win practically is more difficult, at least nowadays, until they would use my nw book lines :) also there is the well known GM draw problem, where sometimes GM seem to have agreed on beforehand to play for a draw. Certain other counting systems/mehtods could change that, but of course it would be better if the game inherently already would be considered (or shown to be) drawish. So imho its a useful thing that i'm trying to show this. But feel free trying to solve it, i dont care\ bye jef
|
| |
Date: 05 Feb 2008 17:02:01
From: David Richerby
Subject: Re: Solving Chess??? (it's a draw!)
|
tin Brown <
|
| |
Date: 05 Feb 2008 17:09:41
From: jefk
Subject: Re: Solving Chess??? (it's a draw!)
|
tin Brown wrote: > Actually Rybka is fantastically strong at beating other engines, but > deprived of its opening book I would not trust it navigate the > intricacies of some openings without guidance from a human. Shredder > and HIARCS seem to do better in this respect in the opening without a > book. yes, Rybka, tends to choose strange moves like 1.Nc3, or Nc6 after d4 d5 Nf3, but the programmer could improve this later of course > > Who is to say that the perfect forced winning line for white does > start with something completely absurd like: 1. h3 1. a3 looks better to me > Only when we have a full depth search result can we be sure. nope, same with checkers, they didnt use a full depth search >> >> claims that there are more positions in chess than atoms in the >> univers. > But that is correct. There are. nah atoms, who cares, i claim there are more quarks in the universe, than there are possible 'real' (originating from legal moves) positions in chess.. 'Proof': well, try to prove its not true, otherwise i hold my claim :) > > Rybka tends to have a very flat evaluation of the initial opening > phase of the game. yes, but most importantly, the initial 'advantage' of white can not be increased during the stage of opening theory. on the contrary, with perfect play black will gradually achieve an equal game > Shredder and HIARCS tend to see more of the structure albeit in a > slightly exaggerated way. yes, but also in this case black can compensate the white initiative >> To Solve Chess we need to search till atleast 100 depth. that's *your* opinion, but with the endgame table bases already a lot of endgame moves will be accounted for > Cataloguing the opening tree within 50cp of equality might make an > interesting net distributed computational project. there seems to exist a project in this direction, but lately unfortunately there doesn't seem to be much progress: http://www.chessbrain.net/ regards, jef
|
| | |
Date: 05 Feb 2008 17:03:21
From: David Richerby
Subject: Re: Solving Chess??? (it's a draw!)
|
jefk <[email protected] > wrote: > yes, but most importantly, the initial 'advantage' of white can not > be increased during the stage of opening theory. on the contrary, > with perfect play black will gradually achieve an equal game Objection. The witness is speculating. Dave. -- David Richerby Natural Impossible Goldfish (TM): www.chiark.greenend.org.uk/~davidr/ it's like a fish but it can't exist and it's completely natural!
|
| | | |
Date: 05 Feb 2008 18:21:15
From: Andy Walker
Subject: Re: Solving Chess??? (it's a draw!)
|
In article <Pfp*[email protected] >, David Richerby <[email protected] > wrote: >jefk <[email protected]> wrote: >> yes, but most importantly, the initial 'advantage' of white can not >> be increased during the stage of opening theory. on the contrary, >> with perfect play black will gradually achieve an equal game >Objection. The witness is speculating. Stronger objection. He is plain wrong. With perfect play, (a) if chess is a draw then White has no initial advantage at all, and Black already has an equal game; or (b) if chess is a White win, then no play by Black can change that [assuming that "perfect play" applies to both sides]; or (c) if chess is [surprisingly] a Black win, then no play by White can change that, and Black has a better-than-equal game. -- Andy Walker Nottingham
|
| | | | |
Date: 05 Feb 2008 21:46:47
From: jefk
Subject: Re: Solving Chess??? (it's a draw!)
|
Andy Walker wrote: > Stronger objection. He is plain wrong. With perfect play, > (a) if chess is a draw then White has no initial advantage at all, > in principle, correct, but in practice moves like 1.e4 and 1.d4 have a small advantage, whereas moves like 1.Nh3, 1.f3, etc have a disadvantage for white. Also you (and Richerby) continuously are disregarding conventional chess theory, where positional considerations/knowledge has grown throughout centuries by chess masters like Steinitz, Nimzowitch, and of course later the more modern super GM's. Besides that we have accumulated knowledge in computer chess, for example it's quite common of course to increase the value of a pawn, the further it has progressed until the other side of the board, threatening to change into a queen, or that a knight should not be in a corner, but better somewhere in the center, a simple concept of mobility. Now from strict game theory you guys might be right to advocate a black swan like a forced win for white, but complicated games like chess and go (not even talking about life itself) are almost as complex as a language imho, and maybe can better not be judged only from a mathematical point of view (in this respect i agree with Kasparov). The limitations of a mathematical approach, whether its in complicated mathematics/physics, life, physics, or maybe also in the area of complicated game theory (ie theory of complicated games, like eg 'Call of duty' (i you know it), depending on levels of course), are described for example in: http://en.wikipedia.org/wiki/Quasi-empiricism_in_mathematics Now back to chess, i agree with Richerby that playing one game with him would be any proof, so lets make another thought experiment: you probably know about socalled the -quite modern- 'free style' computer chess events, where come amateurs with strong computers/engines & extensive opening theory were able to beat some GM's; well these events have rigid time control, but if such an event would be organized with much longer time controls, eg a free-style sort of correspondence chess tournament (similar to conventional correspondence chess where everything is allowed) but then even with *longer* time controls (eg one week/move), then after a few years it most likely will become all draws.. when these guys forget about Sicilian againast e4 at least :) Is that a proof, no of course not, strictly speaking in mathematical terms its not, but its a matter of experience and knowledge - do you know the books by Russian chess expert Shereshevsky for example 'Mastering the endgame , in which he is discussing what type of endgames result from various openings? ; (probably not, as they seem to be sold out) well i have them both, and learned a lot from them. Let me make an even stronger statement about chess, ie *without* a 50 move endgame rule; even then i would doubt whether a win for white could be possible, as after about 25 moves the initial -positional- advantage by white always can be compensated by black;imho. But *if* such a win *would* be possible, its almost certain for me that such a perfect game would be *very* long, hundres, maybe even thousands of moves.. So.. *with* the 50 move endgame rule, there's only one outcome: draw ! (a chess game with thousands of moves is not possible imho with this rule, because of the -very sensible- 3-times repetition rule..) QED.. :) best regard, jef
|
| | | | | |
Date: 06 Feb 2008 00:05:30
From: Andy Walker
Subject: Re: Solving Chess??? (it's a draw!)
|
In article <[email protected] >, jefk <[email protected] > wrote: >> Stronger objection. He is plain wrong. With perfect play, >> (a) if chess is a draw then White has no initial advantage at all, >in principle, correct, but in practice moves like 1.e4 and 1.d4 >have a small advantage, whereas moves like 1.Nh3, 1.f3, etc >have a disadvantage for white. With respect, *you* were the one who told us that with *perfect play* Black could reduce White's initial advantage gradually towards equality. Are you talking about perfect play or about chess as played by humans? >Also you (and Richerby) continuously are disregarding conventional >chess theory, where positional considerations/knowledge has grown >throughout centuries by chess masters like Steinitz, Nimzowitch, >and of course later the more modern super GM's. Of course we aren't. But there is a difference between a game of chess where the players are human, whether beginners or super-GMs, and one where they play "perfect" moves. >Besides that we have accumulated knowledge in computer chess, >for example it's quite common of course to increase the value of a >pawn, the further it has progressed [...] Of course. As imperfect players, we cannot hope to analyse typical positions perfectly through to the final result, so we need a short-hand substitute for this analysis. Positional play, the theories of "small advantages", maxims about control of the centre, about pushing passed pawns, about rooks belonging behind pawns, about rooks being worth more than minor pieces, these are all ways of sumising such substitutes. But with perfect play, a position can only be won, drawn or lost; it can't be "slightly advantageous". Nowadays, we see this in stark form in the ending. The use of tablebases has enabled us to give definitive values -- "perfect play" results -- to many endgame positions, so that recent books now often use quite different conventions about the use of "?" and "!" from the traditional. >Now from strict game theory you guys might be right to advocate a >black swan like a forced win for white, [...] No-one here is "advocating" it; we simply can't rule it out in the present state of knowledge. But there are no "minor wins" in chess, only wins, draws and losses. However important a "slight" advantage may be in practical play, it is of none whatsoever in the theoretical analysis of a position. >Now back to chess, i agree with Richerby that playing one game with >him would be any proof, so lets make another thought experiment: you >probably know about socalled the -quite modern- 'free style' computer >chess events, where come amateurs with strong computers/engines & extensive >opening theory were able to beat some GM's; well these events have >rigid time control, but if such an event would be organized with much >longer time controls, eg a free-style sort of correspondence chess >tournament >(similar to conventional correspondence chess where everything is allowed) >but then even with *longer* time controls (eg one week/move), then >after a few years it most likely will become all draws.. when these >guys forget about Sicilian againast e4 at least :) Unsupported claims again! It seems to me just as likely that expert players backed by extended computer analysis will become ever more adept at squeezing out wins from unpromising positions. To win tournaments of this sort, players will have to take risks -- as they do in super-GM tournaments today -- so that top players will tend to steer towards unfathomable positions and away from clear draws, and towards positions suited to their [and their computers'] talents. But this has little to do with "perfect" play. >Let me make an even stronger statement about chess, ie *without* >a 50 move endgame rule; even then i would doubt whether a win for white >could be possible, as after about 25 moves the initial -positional- >advantage >by white always can be compensated by black;imho. But *if* such >a win *would* be possible, its almost certain for me that such a perfect >game would be *very* long, hundres, maybe even thousands of moves.. Yet more unsupported claims. The endgame tablebases show that in many relatively simple endings a win can be squeezed out in a way that is utterly beyond human comprehension. In the longest variations, the pieces seem to move at random, but the tablebase assures us that *this* is the only way to win, *that* is the most resistant move, and dozens or hundreds of moves later, all becomes clear. Perhaps the initial position is in this same category, and there is an incomprehensible compulsion behind Black's being snared [with best play] into the loss of a pawn by move 47; or perhaps the drawing gin really is quite wide. We won't know in the next few decades; that's as sure as anything is ever sure .... >So.. *with* the 50 move endgame rule, there's only one outcome: draw ! >(a chess game with thousands of moves is not possible imho with >this rule, because of the -very sensible- 3-times repetition rule..) >QED.. Well, it's certainly *possible*; GuyM of these groups has several times posted evidence about what the longest game actually is on the assumption that claimable draws are claimed. It seems very likely that there is no *sensible* game with thousands of moves, unless your measure of "sensible" is how long the players can spin out a game without ever leaving the drawing gins. -- Andy Walker Nottingham
|
| | | | | | |
Date: 06 Feb 2008 12:28:47
From: jefk
Subject: Re: Solving Chess??? (it's a draw!)
|
hello Andy Walker wrote > Of course we aren't. But there is a difference between a > game of chess where the players are human, whether beginners or > super-GMs, and one where they play "perfect" moves. > > ok, i see your point, you (and Richerby) would like to have some 'proof' like a 32 man table base, where immediately the perfect move is given; well i can agree of course that i can't supply such 'proof' .. nevertheless i've given many arguments why i believe (even convinced) that chess is a draw, based on a lot of work/analysis. For example, of course i didn't simply use Rybka in the early stages of the opening, i main use it in deep lines, eg. >40ply, to find counterplans for black (Rybka is not only tactical i think, but also often finds the right plan, at least it seems to do exactly that, based on a positional knowledge). Another thing why Rybka won so many tournaments, not only because its engine, but also because of the Noomen book. He also has done deep analysis, and i've seen his recent articles in the magazine of the Dutch computer chess association (he also described how he participated in a Chessbase organized 'freestyle'event); Yes, people can also choose to go for sharp lines, for black imho its risky to go for Sicilian against a very well prepared opponent, and for white, ah well, you can choose lines like the Danish gambit or so, but a well prepared opponent/strong computer/engine will also manage to squeeze out a draw against sharp gambit lines; or UCO openings.. And again, from gutfeeling (no proof, i admit) i'm convinced that in case of 'perfect play' -if we ever find it- white would not go for gambit lines, except maybe solid pseudo gambits like the QGD (but then i prefer Queens Indian for black) So analysis/bookmaking *is* important, but if i then have managed to find equalizing move in all (=sufficicient) cases for black, then at least i have quite some empirical backup to support my claims; i suppose this will obvious coming years.. (i dont tend to hide my analysis/novelties as the GM's); no it won't be world shocking, but a gradual process, at least that's what i foresee. >> in principle, correct, but in practice moves like 1.e4 and 1.d4 >> have a small advantage, > With respect, *you* were the one who told us that with > *perfect play* Black could reduce White's initial advantage > yes, but against move like 1.e4 its more difficult (less suitable moves) than against 1.f3 > [with best play] into the loss of a pawn by move 47; or perhaps > the drawing gin really is quite wide. We won't know in the next > few decades; that's as sure as anything is ever sure .... > > probably it will take more than a few decades, even for the 'weak solution'; he initiative to start with CAP II was not taken up by enough people, so it was abandoned. And as GuyM pointed out, generated 8+ tablebases will be a huge effort, so even when we take Moore's law into account it most likely will take more than a few decades. or much longer, if nobody bothers to do such an amount of work.. best regards jef
|
| | | | | | | |
Date: 06 Feb 2008 18:10:55
From: Andy Walker
Subject: Re: Solving Chess??? (it's a draw!)
|
In article <[email protected] >, jefk <[email protected]> wrote: >> Of course we aren't. But there is a difference between a >> game of chess where the players are human, whether beginners or >> super-GMs, and one where they play "perfect" moves. >ok, i see your point, you (and Richerby) would like to have some 'proof' >like a 32 man table base, where immediately the perfect move is given; Whoa! You seem determined to show us that you *don't* see the point. "Perfect play" was *your* description of the way in which Black reduced White's initial advantage to equality, not something that I [and perhaps also DavidR] would like proof of. That's not to say that I wouldn't be interested in such a proof, if it happened to exist in the real world; just that it's not going to happen Any Time Soon. My only point is that in the world of perfect play, there is no such thing in chess as a slight advantage, there are only wins, draws and losses. This is different from, for example, ludo or backgammon, where "perfect play" would have probabilistic outcomes, and it might make sense to describe, say, a 51% chance of winning as a "slight advantage". *Im*perfect play also has this probabilistic content, in which context different levels of advantage make sense. [...] >And again, from gutfeeling (no proof, i admit) i'm convinced that >in case of 'perfect play' -if we ever find it- white would not >go for gambit lines, except maybe solid pseudo gambits >like the QGD (but then i prefer Queens Indian for black) In the case of "perfect play", then a gambit line *is* perfect play if it preserves the initial value [such as "draw"] and is a blunder otherwise. If perfect chess is indeed a draw, and if the "drawing gin" is not very slim indeed, then the likelihood is that all sorts of utter rubbish is "perfect" -- a game starting 1 c3 Nc6 2 Qb3 g5 ... seems just as "good" as the latest wrinkle in the Ruy. If we're talking more about practical play in the presence of extremely strong players, then Bronstein, for one, would not have agreed with you. Gambits remain a useful weapon against the computer, given a gambiteer well-versed in anti-computer play. The point is that the pawn is irrelevant to the ultimate outcome of the game, but meanwhile it disrupts the computer's evaluation function. This is even more the case with material imbalances such as RRN vs RBB. >>> in principle, correct, but in practice moves like 1.e4 and 1.d4 >>> have a small advantage, >> With respect, *you* were the one who told us that with >> *perfect play* Black could reduce White's initial advantage >yes, but against move like 1.e4 its more difficult (less suitable moves) >than against 1.f3 As DavidR has commented, this is an important consideration for imperfect human play, irrelevant with perfect play. > [...] And as GuyM pointed out, generated 8+ tablebases >will be a huge effort, so even when we take Moore's law into account >it most likely will take more than a few decades. or much longer, >if nobody bothers to do such an amount of work.. Moore's Law, while it lasts, gives us an extra piece every decade, so we should see the 32-man tablebase in 2260 or so. That is, of course, unless the Law breaks down before that! *Slightly* more realistically, if we ever got to somewhere around the 24-man tablebases, then some opening lines would lead directly into the tablebases [and so be solved]. The effect would be that players would have to avoid exchanges in order to keep the game "interesting", and games would tend to consist of opening "tableaux", with pieces set out by both players, followed by a very sharp period of threatened exchanges into won/drawn positions, and a "mop-up" phase of using the tablebases. -- Andy Walker Nottingham
|
| | | | | | | | |
Date: 06 Feb 2008 22:46:39
From: David Richerby
Subject: Re: Solving Chess??? (it's a draw!)
|
Andy Walker <[email protected] > wrote: > *Slightly* more realistically, if we ever got to somewhere > around the 24-man tablebases, then some opening lines would lead > directly into the tablebases [and so be solved]. Note that this is exactly how Schaeffer's team solved checkers: they built n-man tablebases for some value of n that corresponds to `most of the pieces on the board' and then used exhaustive search up to the point where the tablebases provided terminal evaluations. > The effect would be that players would have to avoid exchanges in > order to keep the game "interesting", and games would tend to > consist of opening "tableaux", with pieces set out by both players, > followed by a very sharp period of threatened exchanges into > won/drawn positions, and a "mop-up" phase of using the tablebases. Well, games between computers would. Human beings would be unable to memorize or understand the tablebases so they wouldn't help us much. Dave. -- David Richerby Aquatic Hilarious Spoon (TM): it's www.chiark.greenend.org.uk/~davidr/ like a piece of cutlery but it's a bundle of laughs and it lives in the sea!
|
| | | | | | | | |
Date: 06 Feb 2008 19:35:03
From: jefk
Subject: Re: Solving Chess??? (it's a draw!)
|
well Andy (actually i prefer Johnny) and David, i'm not interested in a 'perfect' discussion about mathematical logic and philosophy of science, only wanted to show that chess is a draw (it is, btw). And there are various ways for black to maintain a draw; whereby i *assume* that white, trying to win will aim for a long game, whereby black, trying to equalize, will try to get a short -drawish- game. Whether its fundamentally/mathematically speaking 'perfect play' or not if black still would manage to get a draw, but in much more moves than the minimum, i dont care, so i'm out of here, have fun :) cheers, maybe until later, Jef PS yep the 50 move endgame draw rule is stupid in computer chess, but then i'm repeating myself :)
|
| | | | | | | | | |
Date: 07 Feb 2008 10:01:56
From: NOSPAM
Subject: Re: Solving Chess??? (it's a draw!)
|
jefk <[email protected] > shat: >well Andy (actually i prefer Johnny) and David, >i'm not interested in a 'perfect' discussion about >mathematical logic and philosophy of science, Then don't make claims of a mathematical/logical nature that you can't prove. >only wanted to show that chess is a draw (it is, btw). Liar. >i dont care, so i'm out of here, have fun :) Don't let the door hit your ass on the way out, loser.
|
| | | | | | | |
Date: 06 Feb 2008 12:58:25
From: David Richerby
Subject: Re: Solving Chess??? (it's a draw!)
|
jefk <[email protected] > wrote: > Andy Walker wrote >> Of course we aren't. But there is a difference between a game of >> chess where the players are human, whether beginners or super-GMs, >> and one where they play "perfect" moves. > > ok, i see your point, you (and Richerby) would like to have some > 'proof' like a 32 man table base, where immediately the perfect move > is given; well i can agree of course that i can't supply such > 'proof' ... A proof is necessary in order to claim that chess is solved. (Inci- dentally, a full 32-man tablebase would constitute strong solution of chess, since it would indicate how to reach the game-theoretic outcome of every possible position.) Empirical arguments like yours are sufficient for empirical conclusions but not the conclusions you were trying to draw. I agree, even in the absence of a proof, that chess is likely to be a draw with perfect play. But this does not constitute `solving chess'. > nevertheless i've given many arguments why i believe (even > convinced) that chess is a draw, based on a lot of work/analysis. Yes. But you were claiming that these arguments constituted a solution of chess, which they do not. I agree that most of them are evidence that chess probably is a draw. >>> in principle, correct, but in practice moves like 1.e4 and 1.d4 >>> have a small advantage, >> With respect, *you* were the one who told us that with >> *perfect play* Black could reduce White's initial advantage > > yes, but against move like 1.e4 its more difficult (less suitable > moves) than against 1.f3 Agreed. And, for this reason, 1.e4 is, practically, a much better move than 1.f3. But it's still possible that 1.f3 is the only winning move for White, even though he has to walk very carefully to avoid losing. Dave. -- David Richerby Flammable Cat (TM): it's like a cat www.chiark.greenend.org.uk/~davidr/ but it burns really easily!
|
| | | | | | |
Date: 06 Feb 2008 07:47:50
From: Guy Macon
Subject: Re: Solving Chess -Guy Macon
|
Andy Walker wrote: >there is a difference between a game of chess where the >players are human, whether beginners or super-GMs, and >one where they play "perfect" moves. There is a website at Bell Labs that allows us mere humans to experience what playing against a perfect player is like. http://plan9.bell-labs.com/who/ken/chesseg.html The page pits you against a 6-man tablesbase and thus will play perfectly -- but only from the positions available on the page. If White always picks the lowest number, and black always picks the highest, that's perfect play. Or you can hide the numbers until you have decided what to do by moving the browser window so they are off screen, and thus play your skills against those of a perfect player. I like to try to predict what the reply will be to each of my moves -- which at least for me gets easier to do as it gets closer to the end -- and to stop and study the position when I see that my choice isn't close to being the best move available. Humbling, but instructive. -- Guy Macon <http://www.guymacon.com/ >
|
| | | | | | | |
Date: 06 Feb 2008 12:01:54
From: David Richerby
Subject: Re: Solving Chess -Guy Macon
|
Guy Macon <http://www.guymacon.com/ > wrote: > There is a website at Bell Labs that allows us mere humans > to experience what playing against a perfect player is like. > > http://plan9.bell-labs.com/who/ken/chesseg.html > > The page pits you against a 6-man tablesbase and thus will > play perfectly -- but only from the positions available on > the page. > > If White always picks the lowest number, and black always > picks the highest, that's perfect play. Not quite. There is no law of chess that says that shorter wins are preferable. So any move that does not change the outcome of the game is `perfect'. In particular, if one side has a forced win, any move the opponent could make is just as good or bad as any other. Now, for practical play, the player with an advantage wants the game over as quickly as possible, to minimize the chances of mistakes, and the player at a disadvantage wants to drag the game out for as long as possible. Dave. -- David Richerby Indelible Car (TM): it's like a www.chiark.greenend.org.uk/~davidr/ high-performance luxury car but it can't be erased!
|
|
Date: 04 Feb 2008 23:24:19
From: Sanny
Subject: Re: Master Level/ Advance Level will play perfect.
|
> > Here I said "I think" Which means what I think about something. It > > does not mean I am going to bet for it just because I think of it. > > > Here, I am proposing something to have more information so that I can > > judge whether I was Correct. > > =A0 Sanny, you already have oodles of information showing that your > program plays nowhere near perfectly on any of its levels. It is poor- > to-mediocre at best, atrocious at worst. You are quite clearly NOT > correct. > > =A0 So, you are either a liar or a fool. Which is it? I think you have a bit Problem in understanding things. I am neither liar nor a fool. You are doing unnecesary arguments on a thing that do not matter much. If you had beaten Master Level then it does not mean it cannot improve further . It is improving every week. Earlier everyone used to play with Master level and Win. Now most of them are playing with Beginner & Easy Level. Why? The time taken by Master has dropped down from earlier 1 hour to 10 Minutes now. Still people are not willing to play with Master level as they are afraid of loosing the game. Now I find many games where Beginner Level has Won. And I find that everyones rating is going down. If the improvement continues as before the day is not far when It will play as good as Crafty/ Fritz/ Rybka and other Strong Chess Programs. GetClub Game is made just to win against Human Players, I have no motive of having it play with other Commercial programs. As it is made just for fun and not for competing with other Commercial Chess Programs. Bye Sanny Play Chess at: http://www.GetClub.com/Chess.html
|
|
Date: 03 Feb 2008 12:37:58
From:
Subject: Re: Master Level/ Advance Level will play perfect.
|
On Feb 3, 1:22=A0pm, Sanny <[email protected] > wrote: > > > Just as I think Master Level & > > > Advance Level play perfect games. > > > =A0 See? You're still doing it, after admitting you have no basis > > whatever for such a claim! > > A sentence should be read properly. > > I said "Just as I think Master Level & Advance Level play perfect > games". > > Here I said "I think" Which means what I think about something. It > does not mean I am going to bet for it just because I think of it. > > Here, I am proposing something to have more information so that I can > judge whether I was Correct. Sanny, you already have oodles of information showing that your program plays nowhere near perfectly on any of its levels. It is poor- to-mediocre at best, atrocious at worst. You are quite clearly NOT correct. So, you are either a liar or a fool. Which is it?
|
|
Date: 03 Feb 2008 10:22:01
From: Sanny
Subject: Re: Master Level/ Advance Level will play perfect.
|
> > > Just as I think Master Level & > > Advance Level play perfect games. > > =A0 See? You're still doing it, after admitting you have no basis > whatever for such a claim! A sentence should be read properly. I said "Just as I think Master Level & Advance Level play perfect games". Here I said "I think" Which means what I think about something. It does not mean I am going to bet for it just because I think of it. Here, I am proposing something to have more information so that I can judge whether I was Correct. "I think today it will rain" It just means I have a belief that it will happen. There are chances that I may be wrong. But Still I am thinking that way. Because the Probablity is more than 50%. I am proposing something to have more information so that I can judge whether I was Correct. I suppose that clears the objective. Bye Sanny Play Chess at: http://www.GetClub.com/Chess.html
|
|
Date:
From: Martin Brown
Subject: Re: Solving Chess
|
|
| |
Date: 04 Feb 2008 18:55:58
From: jefk
Subject: Re: Solving Chess
|
tin Brown wrote > I don't think that at all. If the game is won for white then there is > at least one perfect game where no matter what black does white can > force a win. theoretical possible, but highly unlikely, if not completely impossible > > In real play yes, but in game theory the final outcome of a > deterministic game like chess with perfect play is one of win, draw or > lose for white. We just don't know which one it is. well after some comprehensive analysis with Ryba i now strongly suspect it's a draw.. so computer chess games probably will become boring in future, eg. Hydra vs Rybka, if Rybka would play on a strong multicpu, than the outcome should be a draw. so you could indeed say that normal chess is 'dead' as the late R.J. Fischer claimed, probably also to promote his random Fischer (or 360) chess; maybe in computer chess it also would be interesting to abolish the 50 move draw rule, as with the 6, and later 7 man endgame bases we could get interesting results (?); also in random/shuffle chess. best regards jef http://superchess.blogspot.com
|
|
Date: 29 Jan 2008 17:39:27
From: pialogue
Subject: Re: Solving Chess
|
First off, the "Sanny/ttk5" exchange appears to be SPAMBOT based and off-topic. Is there any way to delete those messages or at least ban Sanny from posting here? Thank you VERY much for introducing me to the concept of "zugzwang" as I had never previously considered the possibility that chess was an inherently flawed game from the very beginning. I feel like I did when someone introduced me to the concept of oligarchy relative to the concept of politics. ;o) Or perhaps the so-called flaw was actually a form of "trap" that you had to be sure to prepare for and avoid ahead of time? I guess that is the end result either way, a bummer dynamic for the naive or lesser- achieved person? I also get the feeling from the responses above as well as those sent to me privately that the concept of "perfect" is relative depending upon a person's personal opinion or frame of perspective. In other words, in my opinion, there are so many different possibilities for very good plays in the opening and middle games such that the only real "perfect" result would be the game ending in a draw period. A perfect game of defense for black would be to counter or block white's initial theoretical advantage and black's ability to lure white off of a previously-planned successful path. Now obviously I am talking about a perfect game by BOTH players. A perfect game for either side obviously results in a win and I agree that a good endgame database, very soon to be a "totalgame" database, is essential. Thanks!
|
|
Date: 29 Jan 2008 14:07:49
From: help bot
Subject: Re: Solving Chess
|
On Jan 28, 1:22 pm, pialogue <[email protected] > wrote: > So, what are your thoughts on "Solving Chess"? Instead of focusing on draws or perfect defense, a better approach would be to focus on the idea of endgame table-bases. The reason is simple: no matter how deeply you might *calculate* defenses to various and sundry attacks, you could never be certain that there is not a deeper strategy of attack which has not been anticipated, which can prevail. A simplified example of this is where, for many decades, grandmasters preached that certain positions could not be won-- only to die and have their great-great-grandchildren discover they were dead wrong. The strategy was beyond their grasp, but real nonetheless. With the endgame table-bases, no such errors creep in. Sure, there will always be people who maintain that there are more data than can be "held" in the molecules of the "known universe" or whatever, but we've already seen their kind get trampled under- foot before, under the heels of technology and its relentless advance. These small- minded folks remind me of the old days, of back when nearly all the pundits screamed that no chess program would *ever* defeat a master... later changed to "grandmaster", for the obvious reasons. I say that these endgame table-bases *can* be beefed up more and more, until they eventually reach a point where they meet super-engines somewhere in mid game. Right now, Best Buy sells hard drives which can hold more than the nay- sayers said could ever be held, and the price is within reach of ordinary folks. Ten years hence, I expect BB to have been bought out or taken over, but Wal-t, Wal-t will then likely be selling drives which wirelessly connect to our Apple computers (yes, APPL) and which can hold 7, 8, or maybe even 9-man TBs with no trouble at all. Given sufficient time, chess will be solved-- at least to the extent that man will know the definitive answers to such questions as "was Bobby Fischer's Bxh2 a losing move, or could he still have drawn?" And chess engines will be so strong that the question of \winning will be moot; the goal of the world's top GMs will be to somehow squeak out a draw. There will necessarily be a reversion to the old days, to the days of Knight or Rook odds. Discussions will center around just how far can a chess program go, in terms of contempt factors; 2 pawns? Three? Even more? -- help bot
|
|
Date: 29 Jan 2008 10:55:21
From:
Subject: Re: Master Level/ Advance Level will play perfect.
|
On Jan 29, 1:09=A0pm, Sanny <[email protected] > wrote: > On Jan 29, 11:03=A0pm, [email protected] wrote: > > > On Jan 29, 12:48=A0pm, Sanny <[email protected]> wrote: > > > > I am just guessing it may be correct or not. > > > =A0 So then, you really had no basis for saying "Master level will play > > perfect game." In other words, Sanny, you were lying. > > Guessing something does not mean lying. Sanny, the plain fact is that you claimed your program was perfect. You had absolutely no basis for that -- not even the slightest grounds for even guessing that it might be. On the contrary, you have on file hundreds of games proving that it is very imperfect. > Just as I think Master Level & > Advance Level play perfect games. See? You're still doing it, after admitting you have no basis whatever for such a claim! As we say here in the USA, you are talking through your ass.
|
|
Date: 29 Jan 2008 10:09:18
From: Sanny
Subject: Re: Master Level/ Advance Level will play perfect.
|
On Jan 29, 11:03=A0pm, [email protected] wrote: > On Jan 29, 12:48=A0pm, Sanny <[email protected]> wrote: > > > > > I am just guessing it may be correct or not. > > =A0 So then, you really had no basis for saying "Master level will play > perfect game." In other words, Sanny, you were lying. Guessing something does not mean lying. There are many things that we do not know for certain and we guess them. Say after an Exam someone asks you how was the exam, I say It was good I feel I will get 60% ks. But then I get 40% ks. It do not mean I was lying. In Chess who will win can never be said before. There is always scope of improvvement. Even deep blue on a supercomputer was beatable. Even Rybka & Fritz get beaten in some games. But they win most of the time. We can say Fritz & Rybka as an Example of Perfect Game. Just as I think Master Level & Advance Level play perfect games. But that do not mean they will win always. Bye Sanny Play Chess at: http://www.GetClub.com/Chess.html
|
|
Date: 29 Jan 2008 10:03:34
From:
Subject: Re: Master Level/ Advance Level will play perfect.
|
On Jan 29, 12:48=A0pm, Sanny <[email protected] > wrote: > > I am just guessing it may be correct or not. So then, you really had no basis for saying "Master level will play perfect game." In other words, Sanny, you were lying.
|
|
Date: 29 Jan 2008 09:48:36
From: Sanny
Subject: Re: Master Level/ Advance Level will play perfect.
|
On Jan 29, 10:39=A0pm, [email protected] wrote: > On Jan 29, 12:32=A0pm, Sanny <[email protected]> wrote: > > > > > > > On Jan 29, 10:16=A0pm, [email protected] wrote: > > > > On Jan 29, 11:31=A0am, Sanny <[email protected]> wrote: > > > > > > So, what are your thoughts on "SolvingChess"? > > > > > > Any thoughts on the bestchessgame ever played by either human or > > > > > computer would also be greatly appreciated. > > > > > Master Level/ Advance Level atGetClubChesswill play perfect Game as > > > > they will think of 10-20 min/move and play the best game. > > > > =A0 Utter nonsense. Apparently Sanny doesn't even bother to check his > > > own records. Looking through them, I found this game, played just 5 > > > days ago. His "master" level is playing Black; readers may judge for > > > themselves how "perfectly": > > > > 1. e4 c6 2. d4 d5 3. exd5 cxd5 4. c4 Nf6 5. Nc3 dxc4 6. Bxc4 e6 7. Nf3= > > > Nc6 8. O-O h6 9. d5 exd5 10. Re1+ Be6 11. Nxd5 Bc5?? -- Missing an > > > elementary tactical sequence. 12. Nxf6+ Qxf6 13. Bxe6 fxe6 14. Qd5 Be7= > > > 15. Rxe6 -- Winning a pawn and retaining the initiative. 15...Qf7 16. > > > Qe4 O-O? 17. Bxh6 -- Winning another pawn and leaving Black's king > > > highly vulnerable. The game is effectively over. > > > =A0 Sanny's "perfect" program was busted after only a dozen moves. It > > > finally gave up when it had one pawn against two queens. > > > > > Such games will always end in a Draw. > > > > =A0 Yeah, right. And water always flows uphill. > > > Advance Level !!!! And The game was improved 2 days back So Now even > > Master level will play perfect game. > > =A0 Absolute nonsense, Sanny. Would you care to wager, say, $100,000 on > your boast? If I'm going to waste days of my life waiting for your > inept program to make its mediocre-to-lousy moves, I must be > reasonably compensated.- Hide quoted text - > > - Show quoted text - I am just guessing it may be correct or not. Game was improved 2 days back. So Now Master Level will play much strong moves. I am saying after seeing the recorded games. Today PatzerChess took 82 Moves to win Beginner Level. While other guy beat it in just 40 Moves. But I can clearly see that now more and more people are loosing at GetClub. Earlier everyone was playing with Master Level Now no one dares to play with it as Master level will win most of the time. Bye Sanny Play Chess at: http://www.GetClub.com/Chess.html
|
|
Date: 29 Jan 2008 09:39:52
From:
Subject: Re: Master Level/ Advance Level will play perfect.
|
On Jan 29, 12:32=A0pm, Sanny <[email protected] > wrote: > On Jan 29, 10:16=A0pm, [email protected] wrote: > > > > > > > On Jan 29, 11:31=A0am, Sanny <[email protected]> wrote: > > > > > So, what are your thoughts on "Solving Chess"? > > > > > Any thoughts on the best chess game ever played by either human or > > > > computer would also be greatly appreciated. > > > > Master Level/ Advance Level at GetClub Chess will play perfect Game as= > > > they will think of 10-20 min/move and play the best game. > > > =A0 Utter nonsense. Apparently Sanny doesn't even bother to check his > > own records. Looking through them, I found this game, played just 5 > > days ago. His "master" level is playing Black; readers may judge for > > themselves how "perfectly": > > > 1. e4 c6 2. d4 d5 3. exd5 cxd5 4. c4 Nf6 5. Nc3 dxc4 6. Bxc4 e6 7. Nf3 > > Nc6 8. O-O h6 9. d5 exd5 10. Re1+ Be6 11. Nxd5 Bc5?? -- Missing an > > elementary tactical sequence. 12. Nxf6+ Qxf6 13. Bxe6 fxe6 14. Qd5 Be7 > > 15. Rxe6 -- Winning a pawn and retaining the initiative. 15...Qf7 16. > > Qe4 O-O? 17. Bxh6 -- Winning another pawn and leaving Black's king > > highly vulnerable. The game is effectively over. > > =A0 Sanny's "perfect" program was busted after only a dozen moves. It > > finally gave up when it had one pawn against two queens. > > > > Such games will always end in a Draw. > > > =A0 Yeah, right. And water always flows uphill. > > Advance Level !!!! And The game was improved 2 days back So Now even > Master level will play perfect game. Absolute nonsense, Sanny. Would you care to wager, say, $100,000 on your boast? If I'm going to waste days of my life waiting for your inept program to make its mediocre-to-lousy moves, I must be reasonably compensated.
|
|
Date: 29 Jan 2008 09:32:28
From: Sanny
Subject: Re: Master Level/ Advance Level will play perfect.
|
On Jan 29, 10:16=A0pm, [email protected] wrote: > On Jan 29, 11:31=A0am, Sanny <[email protected]> wrote: > > > > So, what are your thoughts on "Solving Chess"? > > > > Any thoughts on the best chess game ever played by either human or > > > computer would also be greatly appreciated. > > > Master Level/ Advance Level at GetClub Chess will play perfect Game as > > they will think of 10-20 min/move and play the best game. > > =A0 Utter nonsense. Apparently Sanny doesn't even bother to check his > own records. Looking through them, I found this game, played just 5 > days ago. His "master" level is playing Black; readers may judge for > themselves how "perfectly": > > 1. e4 c6 2. d4 d5 3. exd5 cxd5 4. c4 Nf6 5. Nc3 dxc4 6. Bxc4 e6 7. Nf3 > Nc6 8. O-O h6 9. d5 exd5 10. Re1+ Be6 11. Nxd5 Bc5?? -- Missing an > elementary tactical sequence. 12. Nxf6+ Qxf6 13. Bxe6 fxe6 14. Qd5 Be7 > 15. Rxe6 -- Winning a pawn and retaining the initiative. 15...Qf7 16. > Qe4 O-O? 17. Bxh6 -- Winning another pawn and leaving Black's king > highly vulnerable. The game is effectively over. > =A0 Sanny's "perfect" program was busted after only a dozen moves. It > finally gave up when it had one pawn against two queens. > > > Such games will always end in a Draw. > > =A0 Yeah, right. And water always flows uphill. Advance Level !!!! And The game was improved 2 days back So Now even Master level will play perfect game. Bye Sanny Play Chess at: http://www.GetClub.com/Chess.html
|
|
Date: 29 Jan 2008 09:16:35
From:
Subject: Re: Master Level/ Advance Level will play perfect.
|
On Jan 29, 11:31=A0am, Sanny <[email protected] > wrote: > > So, what are your thoughts on "Solving Chess"? > > > Any thoughts on the best chess game ever played by either human or > > computer would also be greatly appreciated. > > Master Level/ Advance Level at GetClub Chess will play perfect Game as > they will think of 10-20 min/move and play the best game. Utter nonsense. Apparently Sanny doesn't even bother to check his own records. Looking through them, I found this game, played just 5 days ago. His "master" level is playing Black; readers may judge for themselves how "perfectly": 1. e4 c6 2. d4 d5 3. exd5 cxd5 4. c4 Nf6 5. Nc3 dxc4 6. Bxc4 e6 7. Nf3 Nc6 8. O-O h6 9. d5 exd5 10. Re1+ Be6 11. Nxd5 Bc5?? -- Missing an elementary tactical sequence. 12. Nxf6+ Qxf6 13. Bxe6 fxe6 14. Qd5 Be7 15. Rxe6 -- Winning a pawn and retaining the initiative. 15...Qf7 16. Qe4 O-O? 17. Bxh6 -- Winning another pawn and leaving Black's king highly vulnerable. The game is effectively over. Sanny's "perfect" program was busted after only a dozen moves. It finally gave up when it had one pawn against two queens. > Such games will always end in a Draw. Yeah, right. And water always flows uphill.
|
|
Date: 29 Jan 2008 08:31:44
From: Sanny
Subject: Master Level/ Advance Level will play perfect.
|
> So, what are your thoughts on "Solving Chess"? > > Any thoughts on the best chess game ever played by either human or > computer would also be greatly appreciated. Master Level/ Advance Level at GetClub Chess will play perfect Game as they will think of 10-20 min/move and play the best game. Such games will always end in a Draw. As both will take each others pieces and reach a end game which results in draw. Bye Sanny Play Chess at: http://www.GetClub.com/Chess.html
|
|
Date:
From: Martin Brown
Subject: Re: Solving Chess
|
|
|
Date: 28 Jan 2008 13:27:48
From: pialogue
Subject: Re: Solving Chess
|
On Jan 28, 11:17 am, TrekNoid <[email protected] > wrote: > > Would not a perfectly played game of chess NOT > > result in a draw? How else could it end? The ONLY way that someone > > "wins" is if a "mistake" was made, right? Therefore, if someone wins > > then that was NOT a "perfect" game played by both sides. > > I disagree. You can't make that an assumption of a 'perfect' game, > because both side do not move at the same time. White moves first... > and White's goal is *not* to draw... it's to win. So for White to make > the 'perfect' move, Black would be lost from the first one. Well, if that were the case then wouldn't chess ultimately be like playing Russian Roulette with white having the initial choice of specifically selecting the chamber that did "not" have the bullet?
|
|
Date: 28 Jan 2008 13:17:30
From: TrekNoid
Subject: Re: Solving Chess
|
On Jan 28, 12:22 pm, pialogue <[email protected] > wrote: > Would not a perfectly played game of chess NOT > result in a draw? How else could it end? The ONLY way that someone > "wins" is if a "mistake" was made, right? Therefore, if someone wins > then that was NOT a "perfect" game played by both sides. I disagree. You can't make that an assumption of a 'perfect' game, because both side do not move at the same time. White moves first... and White's goal is *not* to draw... it's to win. So for White to make the 'perfect' move, Black would be lost from the first one.
|
|