| Jan-24-06 | | RandomVisitor: This game shows that Rybka is not invincible. 28...Kg6 is about the worst move possible to play, and this marks the beginning of the end (what was the machine thinking? Oh yeah, the machine can't think...) Rybka strands a piece on the edge of the board, then gives up pawns protecting its King. |
|
| Jan-24-06 | | RandomVisitor: Sorry for this long post, but here is a long analysis of this game, done by Rybka, showing how it evaluates each move: 1. d4 Nf6 2. c4 e6 3. Nc3 Bb4 4. Qc2 c5 5. dxc5 O-O 6. a3 Bxc5 7. Nf3 Nc6 8.
Bg5 b6 9. Rd1 Bb7 10. e4 last book move h6 0.09/19 11. Bh4 0.05/17 g5
0.12/17 12. Bg3 0.12/15 Nh5 0.14/18 13. b4 0.14/18 Be7 0.10/16 14. b5
$11 -0.10/19 (♖ybka 1.01 ♗eta 7: 14. Qd2 Nxg3 15. hxg3 Kg7 16. Qe3 f6 17.
Be2 Qc8 18. O-O a5 19. b5 Bc5 20. Qd3 Ne7 $11 0.10/16) 14... Na5 -0.13/17
15. Ne5 -0.10/17 d6 -0.13/17 16. Ng4 -0.13/16 Qc7 $11 0.00/16 (
♖ybka 1.01 ♗eta 7: 16... Rc8 17. Be5 f6 18. Bd4 Qe8 19. Qa2 Nf4 20. g3 Qh5 21.
Nxh6+ Qxh6 22. gxf4 Rf7 23. fxg5 $11 -0.13/16) 17. Nxh6+ $15 -0.26/18 (
♖ybka 1.01 ♗eta 7: 17. e5 Nxc4 18. exd6 Nxd6 19. Be5 f5 20. Qb3 Nf4 21. Bxf4
gxf4 22. Qxe6+ Kh7 23. Qxh6+ Kg8 $11 0.00/16) 17... Kg7 -0.22/16 18. Ng4
-0.26/16 Nxg3 $11 0.00/19 (♖ybka 1.01 ♗eta 7: 18... Rac8 19. Qb2 Kg8 20.
Ne3 Rfd8 21. Be2 Nf4 22. O-O Bf6 23. Rc1 Qc5 24. Bxf4 gxf4 25. Ng4 $15
-0.26/16) 19. hxg3 0.00/19 Rh8 0.00/19 20. Rxh8 0.00/19 Rxh8 0.00/17
21. Nb1 $15 -0.30/16 (♖ybka 1.01 ♗eta 7: 21. Na4 Rc8 22. Ne3 Bf6 23. Be2
Rh8 24. Ng4 Be7 25. Qc3+ f6 26. Bd3 Rc8 27. Ne3 Rh8 $11 0.00/17) 21... Qc5
$11 -0.02/19 (♖ybka 1.01 ♗eta 7: 21... Rh1 22. Kd2 f6 23. Bd3 Rh8 24. Ne3
Qc5 25. Qc3 Kf7 26. Qb2 Rh2 27. Nc3 Kg7 28. g4 $15 -0.30/16) 22. Qc3+
-0.05/19 f6 -0.02/19 23. f3 -0.05/17 Rh1 0.00/19 24. Ne3 -0.08/17 a6
$11 0.08/18 (♖ybka 1.01 ♗eta 7: 24... Kf7 25. Qd3 a6 26. a4 Qb4+ 27. Kf2
Qxa4 28. Be2 Rxd1 29. Bxd1 Qb4 30. Nc3 axb5 31. cxb5 $11 -0.08/17) 25. a4
0.07/17 g4 0.00/18 26. Rd3 0.00/17 gxf3 0.00/15 27. gxf3 0.00/17 Qg5
0.08/17 28. g4 0.08/16 Kg6 $14 0.27/18 |
|
| Jan-24-06 | | RandomVisitor: Part 2 of analysis: (♖ybka 1.01 ♗eta 7: 28... Qf4 29.
Nd2 Qg3+ 30. Kd1 Qe5 31. Kc2 Rh2 32. Qd4 Qxd4 33. Rxd4 Kg6 34. Rd3 Bf8 35. Rc3
$11 0.08/16) 29. Nd2 0.35/16 Qe5 $14 0.51/18 (♖ybka 1.01 ♗eta 7: 29...
Bc8 30. Qd4 Nb7 31. bxa6 Nc5 32. a7 Nxd3+ 33. Qxd3 Bb7 34. e5+ Kh6 35. exd6
Bxd6 36. Nf5+ $14 0.35/16) 30. Qd4 0.52/16 Qxd4 0.66/20 31. Rxd4 0.67/18
Rh8 0.77/18 32. Rd3 0.68/18 Kf7 0.71/18 33. f4 0.63/17 axb5 $16
0.85/20 (♖ybka 1.01 ♗eta 7: 33... Bc8 34. Bg2 Nb7 35. Nb3 Nc5 36. Nxc5 dxc5
37. Ke2 a5 38. Kf3 Rd8 39. Rxd8 Bxd8 40. e5 $14 0.63/17) 34. axb5 0.84/18
Rh2 $16 1.11/21 (♖ybka 1.01 ♗eta 7: 34... Rc8 35. Kf2 Bf8 36. Be2 Rc5 37.
Ra3 Bh6 38. Kf3 Kg6 39. Ra1 Rc8 40. Rh1 Kg7 41. Bd3 $16 0.84/18) 35. f5
1.11/20 Bc8 1.11/20 36. Bg2 1.15/18 Bb7 $18 1.50/20 (♖ybka 1.01 ♗eta 7:
36... e5 37. Kf2 Bb7 38. Nd5 Bd8 39. Ra3 Rh4 40. Kf3 Rh8 41. Nb3 Nxb3 42.
Rxb3 Kg7 43. Kf2 $16 1.15/18) 37. fxe6+ 1.50/21 Kxe6 1.50/20 38. Kf2
1.50/19 Rh8 1.60/19 39. Nd5 1.59/20 Bd8 1.50/19 40. Rc3 1.60/19 Kf7
1.89/22 41. Nf1 1.82/20 Ke6 1.83/21 42. Ng3 1.85/19 Ba8 1.86/21 43.
Nf5 1.85/20 Bb7 $18 2.47/20 (♖ybka 1.01 ♗eta 7: 43... Kd7 44. Rh3 Rxh3
45. Bxh3 Bxd5 46. exd5 Be7 47. Bf1 Nb3 48. Ke3 Nc5 49. Be2 Bf8 50. Bf3 $18
1.85/20) 44. Rd3 2.47/21 Bxd5 2.47/23 45. exd5+ 2.47/23 Kd7 2.47/22
46. Rc3 2.22/20 Be7 2.48/22 47. Bf1 2.48/20 Bf8 2.91/22 48. Re3
2.91/21 Nb7 2.91/20 49. Ra3 2.65/21 Na5 2.69/22 50. Nd4 2.69/20 Rh2+
2.69/21 51. Kg3 2.66/21 Rh1 $18 3.68/20 (♖ybka 1.01 ♗eta 7: 51... Rd2
52. Nc6 Bh6 53. Kf3 Rd1 54. Bd3 Kc7 55. Bf5 Re1 56. Nxa5 bxa5 57. Rxa5 Re3+ 58.
Kg2 $18 2.66/21) 52. Kg2 3.21/21 Rh4 3.40/21 53. Be2 3.27/20 Be7
3.74/22 54. Re3 3.59/20 Bd8 3.57/21 55. Re6 3.20/22 Rh8 3.20/22 56.
Nc6 3.20/21 Nxc6 4.01/24 57. bxc6+ 4.01/26 Kc7 4.01/26 58. Bd3 4.01/24
Rf8 4.01/26 59. Bf5 4.01/24 Rf7 4.97/26 60. Re8 4.96/22 Re7 4.97/23
61. Rf8 4.97/22 Re3 5.59/23 62. Be6 5.60/21 b5 5.59/21 63. Rf7+
5.59/21 Kb6 6.49/20 64. Rb7+ 6.07/19 Kc5 6.56/23 65. c7 6.56/25 Bxc7
6.30/23 66. Rxc7+ 5.77/22 Kb4 6.29/22 67. cxb5 5.29/20 Kxb5 $2 $18
8.35/26 (♖ybka 1.01 ♗eta 7: 67... Ka5 68. Kf2 Rb3 69. Bd7 Rb1 70. Rc6 Rxb5
71. Rc8 Rb6 72. Ke3 Ka6 73. Rc4 Rb3+ 74. Kd4 $18 5.29/20) 68. Rf7 8.35/24
Kc5 8.35/22 69. Rxf6 8.35/24 * |
|
| Jan-24-06 | | csmath: <This game shows that Rybka is not invincible.> It certainly isn't. Here is my game against Rybka, I used mostly Deep Shredder 9: [Event "ICC 20 30"]
[Site "Internet Chess Club"]
[Date "2006.01.07"]
[Round "-"]
[White "McPillen"]
[Black "CSMath"]
[Result "0-1"]
[ICCResult "White checkmated"]
[WhiteElo "2496"]
[BlackElo "2905"]
[Opening "Trompowsky attack (Ruth, Opocenský opening)"]
[ECO "A45"]
[NIC "QP.07"]
[Time "20:35:47"]
[TimeControl "1200+30"]
1. d4 Nf6 2. Bg5 d5 3. Bxf6 exf6 4. e3 c6 5. Bd3 Qb6 6. b3 Bd6 7. Ne2 O-O 8.
c4 dxc4 9. bxc4 c5 10. Nbc3 Nc6 11. O-O Nb4 12. Nd5 Nxd5 13. cxd5 c4 14.
Bxc4 Qc7 15. Rc1 Bxh2+ 16. Kh1 Bd6 17. Bd3 Qa5 18. Qb3 f5 19. Nc3 a6 20. a4
b6 21. Rb1 Rb8 22. e4 fxe4 23. Nxe4 Bf4 24. g3 f5 25. gxf4 fxe4 26. Bxe4 Bd7
27. Rg1 Rf6 28. Qd3 Kh8 29. Rg5 Rh6+ 30. Kg2 Qxa4 31. Rc1 Qb4 32. Bf5 Qd6
33. Bxd7 Qxd7 34. Qf5 Qxf5 35. Rxf5 b5 36. Kf3 b4 37. Rf7 Rhb6 38. Rc6 h5
39. Rcc7 h4 40. Rxg7 Rh6 41. Rg1 b3 42. Rc6 Rh7 43. Rxa6 b2 44. Rb1 h3 45.
Re6 Kg8 46. Ree1 Rb3+ 47. Kg4 h2 48. d6 Kf8 49. f5 h1=Q 50. Rxh1 Rxh1 51.
Rxh1 b1=Q 52. Rxb1 Rxb1 53. Kf4 Re1 54. f6 Re6 55. d7 Rd6 56. Ke5 Rxd7 57.
Ke4 Kf7 58. Ke5 Rd8 59. d5 Rh8 60. d6 Rh1 61. Ke4 Kxf6 62. Kd4 Rd1+ 63. Kc5
Ke5 64. f4+ Kxf4 65. Kc6 Ke5 66. d7 Ke6 67. Kc5 Rxd7 68. Kc4 Rd5 69. Kc3 Kd6
70. Kb3 Kc5 71. Kc3 Kb5 72. Kc2 Kb4 73. Kc1 Kb3 74. Kb1 Rd1# White
checkmated 0-1
Here is another win against Rybka by IM from Egypt. This is actually a demolition: [Event "ICC 15 0"]
[Site "Internet Chess Club"]
[Date "2006.01.15"]
[Round "-"]
[White "B-Amin"]
[Black "tribbles"]
[Result "1-0"]
[ICCResult "Black checkmated"]
[WhiteElo "2651"]
[BlackElo "2939"]
[Opening "Sicilian defense"]
[ECO "B40"]
[NIC "SI.43"]
[Time "06:09:42"]
[TimeControl "900+0"]
1. e4 c5 2. Nf3 e6 3. d3 Nc6 4. g3 d5 5. Nbd2 g6 6. Bg2 Nge7 7. O-O Bg7 8.
Re1 O-O 9. Qe2 b6 10. Nf1 Bb7 11. e5 Qd7 12. h4 Nf5 13. c3 d4 14. c4 Rfe8
15. N1h2 b5 16. b3 bxc4 17. bxc4 h5 18. g4 Nfe7 19. Bg5 Nb4 20. a3 Nbc6 21.
Qd2 Bf8 22. Bf6 Reb8 23. gxh5 Nf5 24. Ng4 Qe8 25. Bg5 Qc8 26. Bf6 Qe8 27.
Qg5 a5 28. Kh2 Na7 29. Re4 Kh7 30. Qf4 Kg8 31. Rg1 Bc6 32. hxg6 fxg6 33. h5
Rb7 34. hxg6 Qxg6 35. Bh3 Kf7 36. Ng5+ Ke8 37. Nxe6 Rh7 38. Nxf8 Kxf8 39.
Bg5 Rxh3+ 40. Kxh3 Nc8 41. Nh6 Qh5+ 42. Kg2 Qg6 43. Qxf5+ Qxf5 44. Nxf5 Ra6
45. Rh1 Bxe4+ 46. dxe4 Rg6 47. f4 Kg8 48. Kf3 d3 49. Ke3 d2 50. Kxd2 Rc6 51.
Ke3 Kf7 52. Rd1 Nb6 53. Nd6+ Kg8 54. f5 Kf8 55. Kf4 Nc8 56. Nxc8 Rxc8 57.
Rd8+ Rxd8 58. Bxd8 a4 59. Bb6 Ke7 60. Bxc5+ Kf7 61. e6+ Ke8 62. f6 Kd8 63.
f7 Kc7 64. f8=Q Kb7 65. e7 Kc7 66. e8=Q Kb7 67. e5 Ka6 68. e6 Ka5 69. e7 Ka6
70. Qd7 Ka5 71. e8=Q Ka6 72. Qb8 Ka5 73. Qda7# ♗lack checkmated 1-0 |
|
| Jan-26-06 | | EricCartman: <RandomVisitor> This game also shows us that Rybka doesn't understand anything about the most important thing in chess: the pawnstructure. During the whole game it's pawnstructure is a mess. This is the reason why Rybka would be no match for Hydra. Take a look at this game. Hydra vs Adams, 2005 Move 11 is brilliant, it weakens the pawnstructer of Adams so badly, that he eventually lost the game (and the whole match). |
|
| Jan-26-06 | | RandomVisitor: <During the whole game it's pawnstructure is a mess> Yes. Perhaps the author of Rybka needs to put more "smarts" in this area. |
|
| Jan-27-06 | | EricCartman: <RandomVisitor> That's a good idea. Rybka would be hell of a lot stronger then |
|
| Feb-03-06 | | sharpnova: how do you know? there's a balance to be struck in any chess program.. if you put "smarts" into pawn structure you're going to have to take smarts out of something else or else have a drastic reduction in nodes/sec etc. think about it. |
|
| Feb-03-06 | | RandomVisitor: <if you put "smarts" into pawn structure you're going to have to take smarts out of something else or else have a drastic reduction in nodes/sec etc. think about it.> Well... Rybka, for example, already has a drastic reduction in nodes per second. It is the slowest engine I have worked with. However, if you have an accurate position analysis, then you can cut back your search tree and make up whatever time you lost being meticulous in your evaluation. But yes, it will cost you something in terms of performance. The big question is weather there is a net gain in ELO performance. I know that Rybka has an extensive network of people who are testing prototype versions, so that if a proposed change results in a weaker program, then the proposed change is not used. |
|
| Feb-03-06 | | RookFile: I don't agree with this discussion regarding pawn structure. Maybe you don't like ....h6 and ...g5 but that happens all the time in these lines in the Nimzo. |
|
| Feb-22-06 | | Everett: I do agree with this discussion that if one would "improve" a specific quality of chess performance, it most likely will effect other parts of its play. In this sense, the "perfect" program may never be created. |
|
| Apr-21-06 | | MVL: Well of course is not perfect but it sure is the best out there forget about fritz shredder or junior not even Hydra wich is only powerfull because it runs on 64 processors but it sucks in many areas,Rybka is the closest program i've had that understands a chess position more human-like,i own a multiprocessor machine and it outplays the Deep shredders deep fritz 8 etc using only one processor while the others use full power make no mistake about it gentleman Rybka will be absolute #1 over everybody (humans and machines) if it keeps improving |
|
| Aug-14-06 | | spirit: complex computer ending |
|
| Jan-06-10 | | zanshin: <RandomVisitor: This game shows that Rybka is not invincible. 28...Kg6 is about the worst move possible to play> I don't know if Rybka has changed that much since 2005. Current analysis of the position: 
click for larger view [+0.11] d=17 28...axb5 29.cxb5 Qf4 30.Nd2 d5 31.exd5 Bb4 32.Qd4 Qg3 33.Kd1 Bxd2 34.Rxd2 Qxf3 35.Kc2 Rxf1 36.Nxf1 Qb3 37.Kc1 Qf3 38.Qd3 Nb3 39.Kc2 Nxd2 40.Nxd2 Qxg4 41.dxe6 Qxe6 42.Qg3 Kf8 43.Qb8 Bc8 (0:26.27) 188730kN Maybe <RV> was right .. she does not like <28...Kg6> so much any more. [+0.55] d=16 28...Kg6 29.Nd2 Qe5 30.Qd4 Qxd4 31.Rxd4 Kf7 32.Kf2 Rh8 33.Rd3 Rc8 34.f4 Bf8 35.f5 (0:13.31) 77376kN But I think much of the problem Rybka had in this game was shown in the shuffling of pieces. She could not "find something to do." |
|
Kibitzing
icon.