Author: James T. Walker
Date: 15:14:29 11/20/98
Go up one level in this thread
On November 19, 1998 at 21:31:56, Howard Exner wrote: >On November 19, 1998 at 09:24:14, James T. Walker wrote: > >>Hello, >>This is game 8 from my match betweem Rebel 10 and CM5500. I am using both >>programs as they "come out of the box" with one exception. I set CM5500 to use >>64 MB for hash tables. It didn't seem fair for Rebel 10 to have 100 MB and >>CM5500 only 1MB. The first 6 games Rebel 10 had the AMD K6-2 350 mhz machine. >>For the last 6 games Rebel 10 will have the PII 333 mhz machine. >>Score to this point: >> >>Game 1 2 3 4 5 6 7 8 Total >>Rebel 10 0 1 .5 .5 1 .5 1 0 4.5 >>CM5500 1 0 .5 .5 0 .5 0 1 3.5 >> >> >>For this game: >>Rebel 10 - PII 333mhz >>CM5500 - AMD K6-2 350 mhz >> >>[Event "40/90"] >>[Site "FL"] >>[Date "1998.11.18"] >>[Round "8"] >>[White "REBEL 10.0"] >>[Black "CM5500"] >>[Result "0-1"] >>[WhiteElo "2552"] >>[BlackElo "2500"] >>[ECO "A50"] >> >>1. d4 Nf6 2. c4 c6 3. Nf3 d5 4. Nc3 dxc4 5. a4 Bf5 6. e3 e6 7. Bxc4 Bb4 >>8. O-O O-O 9. Nh4 Bg4 10. f3 Nd5 11. fxg4 Qxh4 12. Qf3 Nd7 13. e4 N5b6 >>14. Bb3 c5 15. Na2 Ba5 16. Be3 Qe7 17. Rac1 Rac8 18. Qf2 c4 19. Bc2 e5 >>20. d5 Nc5 21. Bxc5 Rxc5 22. Rcd1 Nc8 23. Qf3 Nd6 24. Nc3 Bb6 25. Kh1 Rc7 >>26. Ne2 Qg5 27. a5 Bc5 28. Nc3 Bd4 29. Qg3 Rcc8 30. h3 Qe7 31. Ra1 Rfd8 >>32. Qf3 Rb8 33. Ba4 Qc7 34. Bd1 b5 35. axb6 Rxb6 36. Ra2 Rdb8 37. Na4 Rb4 >>38. Qe2 Qb7 39. Kh2 a5 40. Bc2 Qc7 41. Rb1 Qa7 42. Rba1 Qe7 43. Bd1 Qf6 >>44. g3 Qd8 45. Nc3 R8b7 46. Rc1 Qb6 47. Na4 Qa7 48. Qg2 Rb8 49. Rc2 Rb3 >>50. Re2 Re3 51. Rxe3 Bxe3 52. Nc3 Bd4 53. Bc2 Qb6 54. Nd1 Qb4 55. Qe2 Rb7 >>56. Ba4 Rb6 57. g5 Qc5 58. Bd7 Rb8 59. Bg4 Rb3 60. Qe1 Rb4 61. Bh5 Qb6 >>62. Qe2 Rb3 63. Qe1 g6 64. Bg4 Qb4 65. Qe2 Rd3 66. Nc3 Be3 67. Qf1 Rd2+ >>68. Be2 Bd4 69. Qe1 Rxb2 70. Rxb2 Qxb2 71. Nd1 Qc2 72. Nf2 a4 73. Bd1 Qb2 >>74. Bxa4 Bxf2 75. Qh1 Qc3 76. Qg2 Nxe4 77. h4 Bxg3+ 78. Kh1 Nf2+ 79. Kg1 >>Qe1+ 80. Qf1 Nh3+ 81. Kg2 Nf4+ 82. Qxf4 Bxf4 83. Bd1 Qg3+ 84. Kf1 Be3 85. >>Ke2 Qf2# 0-1 >> >>This was in my opinion one of the more "interesting' games. Frankly, I'm bored >>to tears watching these two play at this time control. It should not be as >>boring replaying them at whatever speed you like. More to follow. > >I'm not sure I'd be able to manually enter in the moves as you are doing. >Having only one computer I do not have that luxury. I do like to play over these >games and so thank you and all those you take the time to play and post them. > >About the game, it was a nice victory by CM5500 with it gradually building up >its position and allowing Rebel little room to squirm out of its constricted >position. One question I've always wanted to ask, and this I think could only >apply to manual testing. If you noted, how often did the opponents in this game >predict the moves made by each other. This would be interesting to know, as >well as how long each move took to make. > >One move of Rebel's I could not produce was move 59. Bg4. On all antigm settings >Rebel never once considers this move. I don't doubt for one minute that Rebel >played this move as it could be the difference in our hash sizes >(I use 60 MB while you had 100). Out of curiousity how does Rebel 10 eval >this move (on my machine Bc6 was the preferred move)? Do you recall if >Rebel was in a big hash hit when it made this move (ie: predicting successfully >Cm5500's moves prior to its own move 59)? > Hello Howard, I went back and put Rebel 10 on infinite to run move 59 again. It took 3:34 to make the move 59.Bg4. It was 9.03 ply and the score was 0.06. It was 59. Bc6 up to that point with a score of 0.04. And yes I find these two programs predict each others moves at a seemingly high rate. I don't document this kind of thing anymore. I found that when I played crafty 16.0 against Rebel they seldom predicted each others moves (In contrast). I played them at Game/30 though. Jim Walker >If you can please post the other game scores.
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.