Author: Uri Blass
Date: 04:08:05 07/24/02
Go up one level in this thread
On July 24, 2002 at 03:04:16, Nolan Denson wrote:
>I was wondering with the speeds of the EGTB can i gain Strenght in play by not
>doing probes. Since i did the following ... Matrix seems to have gain some
>strenght ... It has yet to lose and in the end game it played one of the most
>empressive endings ever. Please check out this game and notice the streak that
>it has now.
>
>44: + 2831 B 2754 PostModernis [ br 3 0] E33 Res Jul 24 02 02:32
>43: + 2817 W 2779 MoonShot [ br 3 0] D52 Mat Jul 24 02 02:26
>42: + 2801 W 2879 torque [ br 3 0] A40 Res Jul 24 02 02:20
>41: = 2780 B 2900 torque [ br 3 0] D02 Rep Jul 24 02 02:06
>40: = 2774 W 2906 torque [ br 3 0] B09 Rep Jul 24 02 02:02
>39: = 2768 B 2764 SteffCrafty [ br 3 0] E59 NM Jul 24 02 01:42
>38: + 2768 W 2764 SteffCrafty [ br 3 0] E47 Mat Jul 24 02 01:36
>37: = 2751 B 2781 SteffCrafty [ br 3 0] A46 Agr Jul 24 02 01:32
>36: = 2749 W 2783 SteffCrafty [ br 3 0] E47 Rep Jul 24 02 00:51
>
>[Result "0-1"]
>[ICCResult "White resigns"]
>[WhiteElo "2754"]
>[BlackElo "2831"]
>[Opening "Nimzo-Indian: classical, 4...Nc6"]
>[ECO "E33"]
>[NIC "NI.24"]
>[Time "02:32:22"]
>[TimeControl "180+0"]
>
>1. d4 Nf6 2. c4 e6 3. Nc3 Bb4 4. Qc2 Nc6 5. e3 O-O 6. Nf3 d5 7. Be2 dxc4 8. Bxc4
>a6 9. O-O Bd6 10. a3 e5 11. Nxe5 Nxe5 12. dxe5 Bxe5 13. Rd1 Qe7 14. Be2 Qc5 15.
>Qd3 Bf5 16. e4 Ng4 17. Bxg4 Bxg4 18. Be3 Qd6 19. Qxd6 Bxd6 20. Rd4 Be5 21. Rd5
>f6 22. f3 Be6 23. Rd2 Rad8 24. Rad1 Rxd2 25. Rxd2 b5 26. g3 Bxc3 27. bxc3 h6 28.
>Bf4 Rc8 29. Kf2 g5 30. Be3 Kf7 31. Bc5 Re8 32. Ke3 h5 33. a4 bxa4 34. Rd4 a3 35.
>Bxa3 g4 36. f4 a5 37. Kd2 Kg7 38. f5 Bf7 39. Kd3 Rb8 40. Rd7 Rb3 41. Be7 a4 42.
>Rxc7 a3 43. e5 fxe5 44. Bd6 a2 45. Bxe5+ Kf8 46. Ra7 Rb1 47. Ra8+ Ke7 48. c4
>Rd1+ 49. Ke2 Rh1 50. Ra7+ Ke8 51. c5 Rxh2+ 52. Kd3 Bd5 53. f6 Rf2 54. Kd4 Rd2+
>55. Ke3 Rc2 56. Bd4 h4 57. gxh4 g3 58. Kf4
>
>[D]4k3/R7/5P2/2Pb4/3B3P/4K1p1/p1r5/8 w - - 0 58
>
>You can see what both programs were thinking from the below eval
>
>PostModernist(C) kibitzes: PM: depth=8 score= +1.00 time=1 nps=115698
>egtbprobes=0 egtbhits=0 PV: 57.gxh4 g3 58.Kf4 g2 59.Ke5 Bf7 60.Kd6 Rd2
>aics% aics% TombRaider(C) kibitzes: ply=9; eval=-0.21; nps=744k; time=2.06;
>cpu=345%; egtb=0
>
>
>g2 59. Ke5 Rd2 60. Bg1 Rd1 61. Bf2 Bf7 62. h5 a1=Q+ 63. Rxa1 Rxa1 64. h6 Kd7 65.
>h7 Rh1 66. Kf4 Rxh7 67. Kg3 Bd5 68. Kf4 Ke6 69. Ke3 Kxf6 70. Kd3 Rh3+ 71. Kc2
>Be4+ 72. Kb2 Ke5 73. Ka2 Rc3 {White resigns} 0-1
>
>I think the use of no EGTB's helped Matrix win this game. It is my belief that
>the extra speed game was more important.
>
>So do fast computers need EGTB's ???
I believe that EGTB can help if you use
them correctly.
I believe that Crafty is using them in a productive way
so matrix with EGTB's should be better than matrix without them.
You can earn speed from tablebases because
the program does not waste time in searching positions
that it knows the result and if the remaining depth
is big enough then probing the tablebases is faster
than searching.
Uri
This page took 0 seconds to execute
Last modified: Thu, 15 Apr 21 08:11:13 -0700
Current Computer Chess Club Forums at Talkchess. This site by Sean Mintz.