Rybka 3 - Crafty 22.1 Single CPU

[Dr.Ex]

In Blitz time controls on single core there was nothing else to be expected. Rybka 2 would most probably achieve a similar result.

[Dr.Wael Deeb]

I agree with you. The new Rybka style is quite aggrssive, much more fun to watch.

This was my first test game with the new Rybka 3. Then on I knew that this was a different Rybka.

[Event "--, Blitz:5'+4""]
[Site "-----"]
[Date "2008.08.07"]
[Round "1"]
[White "Zappa Mexico II"]
[Black "Rybka 3"]
[Result "0-1"]
[ECO "B00"]
[Annotator "0.30;0.10"]
[PlyCount "194"]
[TimeControl "300+4"]

{Intel(R) Core(TM)2 Duo CPU T9300 @ 2.50GHz 0 MHz W=15.2 ply; 1,175kN/s;
Fritz11.ctg B=15.9 ply; 120kN/s; Fritz11.ctg} 1. e4 {
[%eval 0,0] [%emt 0:00:00]} e5 {[%eval 0,0] [%emt 0:00:00]} 2. d4 {
[%eval 0,0] [%emt 0:00:00]} Nc6 {[%eval 0,0] [%emt 0:00:00]} 3. d5 {
[%eval 0,0] [%emt 0:00:00]} Nce7 {[%eval 0,0] [%emt 0:00:00]} 4. Bd3 {
[%eval 0,0] [%emt 0:00:00]} Ng6 {[%eval 0,0] [%emt 0:00:00]} 5. Be3 {
Both last book move [%eval 0,0] [%emt 0:00:00]} Nf6 {
[%eval 10,15] [%emt 0:00:18]} 6. a3 {(Nf3) [%eval 30,13] [%emt 0:00:34]} c6 {
(b6) [%eval -4,13] [%emt 0:00:10]} 7. Nc3 {(c4) [%eval 27,12] [%emt 0:00:17]}
Nh4 {(b5) [%eval 8,12] [%emt 0:00:14]} 8. Bf1 {[%eval 41,13] [%emt 0:00:26]} d6
{(Qa5) [%eval 17,14] [%emt 0:00:21]} 9. h3 {(g3) [%eval 51,13] [%emt 0:00:16]}
Bd7 {(Be7) [%eval 13,14] [%emt 0:00:14]} 10. Nf3 {
(dxc6) [%eval 38,14] [%emt 0:00:29]} Nxf3+ {[%eval 9,14] [%emt 0:00:08]} 11.
Qxf3 {[%eval 42,16] [%emt 0:00:23]} Be7 {(cxd5) [%eval 10,15] [%emt 0:00:18]}
12. g4 {(Bd3) [%eval 42,14] [%emt 0:00:10]} cxd5 {[%eval 14,14] [%emt 0:00:14]}
13. exd5 {[%eval 34,15] [%emt 0:00:28]} h6 {[%eval 11,15] [%emt 0:00:09]} 14.
O-O-O {[%eval 48,15] [%emt 0:00:26]} Rc8 {[%eval 25,14] [%emt 0:00:15]} 15. Qg2
{(Rg1) [%eval 47,14] [%emt 0:00:08]} O-O {[%eval 3,13] [%emt 0:00:13]} 16. f4 {
[%eval 49,13] [%emt 0:00:12]}

[D]2rq1rk1/pp1bbpp1/3p1n1p/3Pp3/5PP1/P1N1B2P/1PP3Q1/2KR1B1R b - - 0 16
Here Rybka plays Rxc3! and leads a nice game. (I don't know how to do a diagram.)


Rxc3 {(exf4) [%eval -32,13] [%emt 0:00:18]} 17.
bxc3 {[%eval 58,8] [%emt 0:00:00]} Qa5 {(exf4) [%eval -15,14] [%emt 0:00:33]}
18. Kb2 {[%eval 32,14] [%emt 0:00:19]} exf4 {
(Rc8) [%eval -55,14] [%emt 0:00:49]} 19. Bxf4 {[%eval -7,12] [%emt 0:00:06]}
Rc8 {[%eval -50,13] [%emt 0:00:06]} 20. c4 {[%eval -7,13] [%emt 0:00:08]} Nxd5
{[%eval -63,13] [%emt 0:00:07]} 21. Rxd5 {[%eval -45,13] [%emt 0:00:12]} Bf6+ {
[%eval -63,14] [%emt 0:00:34]} 22. Kc1 {[%eval -45,13] [%emt 0:00:00]} Qxa3+ {
[%eval -63,12] [%emt 0:00:00]} 23. Kd1 {[%eval -83,15] [%emt 0:00:35]} Qa1+ {
[%eval -63,13] [%emt 0:00:27]} 24. Bc1 {[%eval -56,14] [%emt 0:00:11]} Re8 {
[%eval -29,13] [%emt 0:00:06]} 25. Bd3 {(Be2) [%eval -56,15] [%emt 0:00:10]}
Bc6 {[%eval -119,12] [%emt 0:00:05]} 26. h4 {[%eval -56,15] [%emt 0:00:19]} Bd4
{[%eval -117,12] [%emt 0:00:03]} 27. c3 {[%eval -56,13] [%emt 0:00:00]} Ba4+ {
[%eval -118,13] [%emt 0:00:13]} 28. Bc2 {[%eval -56,14] [%emt 0:00:00]} Be3 {
[%eval -118,12] [%emt 0:00:00]} 29. Rd2 {[%eval -56,13] [%emt 0:00:00]} Bc6 {
[%eval -110,14] [%emt 0:00:13]} 30. Qf1 {[%eval -62,13] [%emt 0:00:07]} Bxh1 {
[%eval -110,13] [%emt 0:00:02]} 31. Qxh1 {[%eval -62,12] [%emt 0:00:00]} Qxc3 {
[%eval -115,14] [%emt 0:00:15]} 32. Qxb7 {[%eval -58,13] [%emt 0:00:11]} Bxd2 {
[%eval -115,13] [%emt 0:00:00]} 33. Bxd2 {[%eval -58,13] [%emt 0:00:00]} Qh3 {
[%eval -118,14] [%emt 0:00:13]} 34. Bf5 {[%eval -50,12] [%emt 0:00:07]} Qf1+ {
[%eval -118,14] [%emt 0:00:11]} 35. Kc2 {[%eval -50,12] [%emt 0:00:00]} Qxc4+ {
[%eval -118,14] [%emt 0:00:09]} 36. Bc3 {[%eval -50,13] [%emt 0:00:19]} Qe2+ {
(Kf8) [%eval -129,14] [%emt 0:00:06]} 37. Kb3 {[%eval -43,13] [%emt 0:00:08]}
g6 {(Qd1+) [%eval -129,14] [%emt 0:00:03]} 38. Bd7 {
[%eval -29,13] [%emt 0:00:06]} Qd1+ {[%eval -129,14] [%emt 0:00:06]} 39. Ka3 {
(Ka2) [%eval -60,15] [%emt 0:00:13]} Qc1+ {[%eval -134,14] [%emt 0:00:03]} 40.
Ka2 {(Qb2) [%eval -69,15] [%emt 0:00:13]} Re2+ {[%eval -151,13] [%emt 0:00:06]}
41. Bb2 {[%eval -63,14] [%emt 0:00:00]} Qc4+ {[%eval -144,14] [%emt 0:00:06]}
42. Qb3 {[%eval -63,13] [%emt 0:00:00]} d5 {
(Re4) [%eval -144,14] [%emt 0:00:02]} 43. Bb5 {[%eval -77,15] [%emt 0:00:07]}
Qxb3+ {[%eval -144,12] [%emt 0:00:00]} 44. Kxb3 {[%eval -77,14] [%emt 0:00:00]}
Re3+ {[%eval -145,16] [%emt 0:00:08]} 45. Kc2 {[%eval -78,15] [%emt 0:00:06]}
Re4 {[%eval -142,16] [%emt 0:00:06]} 46. Bd7 {[%eval -79,15] [%emt 0:00:06]}
Re7 {(a5) [%eval -147,17] [%emt 0:00:06]} 47. Bb5 {
(Ba4) [%eval -82,16] [%emt 0:00:10]} Rc7+ {[%eval -148,16] [%emt 0:00:04]} 48.
Kb1 {[%eval -86,16] [%emt 0:00:04]} Rb7 {(a5) [%eval -150,17] [%emt 0:00:04]}
49. Be2 {[%eval -80,15] [%emt 0:00:03]} a5 {[%eval -149,16] [%emt 0:00:03]} 50.
Bd1 {(Kc2) [%eval -87,15] [%emt 0:00:03]} Rb4 {[%eval -174,16] [%emt 0:00:05]}
51. Kc2 {[%eval -81,16] [%emt 0:00:10]} a4 {[%eval -181,17] [%emt 0:00:04]} 52.
Kc3 {(h5) [%eval -103,15] [%emt 0:00:04]} Rf4 {
(Rc4+) [%eval -184,15] [%emt 0:00:03]} 53. Bc1 {[%eval -111,15] [%emt 0:00:05]}
Rc4+ {[%eval -184,16] [%emt 0:00:03]} 54. Kd2 {[%eval -111,14] [%emt 0:00:00]}
Kg7 {(Rd4+) [%eval -184,16] [%emt 0:00:06]} 55. Bb2+ {
[%eval -128,15] [%emt 0:00:05]} Kf8 {(f6) [%eval -165,16] [%emt 0:00:04]} 56.
Be2 {[%eval -97,14] [%emt 0:00:04]} Rf4 {[%eval -178,17] [%emt 0:00:04]} 57.
Be5 {[%eval -111,14] [%emt 0:00:05]} Re4 {[%eval -178,15] [%emt 0:00:00]} 58.
Bd6+ {[%eval -99,14] [%emt 0:00:05]} Ke8 {[%eval -184,16] [%emt 0:00:02]} 59.
h5 {(Bf3) [%eval -108,16] [%emt 0:00:06]} g5 {
(Kd7) [%eval -184,16] [%emt 0:00:04]} 60. Bf3 {[%eval -101,15] [%emt 0:00:07]}
Rc4 {(Rd4+) [%eval -188,17] [%emt 0:00:04]} 61. Be2 {
[%eval -101,15] [%emt 0:00:03]} Rd4+ {[%eval -188,15] [%emt 0:00:00]} 62. Kc2 {
(Ke3) [%eval -101,15] [%emt 0:00:06]} f6 {[%eval -188,17] [%emt 0:00:02]} 63.
Bb5+ {(Ba3) [%eval -119,15] [%emt 0:00:03]} Kd8 {[%eval -188,19] [%emt 0:00:05]
} 64. Be2 {[%eval -119,15] [%emt 0:00:00]} Kd7 {
(Re4) [%eval -188,19] [%emt 0:00:09]} 65. Bf8 {
(Bc5) [%eval -141,15] [%emt 0:00:03]} Ke6 {[%eval -190,16] [%emt 0:00:04]} 66.
Kc3 {[%eval -139,15] [%emt 0:00:03]} Re4 {[%eval -210,16] [%emt 0:00:07]} 67.
Bd1 {[%eval -144,14] [%emt 0:00:03]} Re3+ {(Kf7) [%eval -241,15] [%emt 0:00:02]
} 68. Kb2 {[%eval -201,15] [%emt 0:00:06]} Rg3 {[%eval -252,16] [%emt 0:00:05]}
69. Bxa4 {[%eval -175,15] [%emt 0:00:04]} Rxg4 {[%eval -252,14] [%emt 0:00:00]}
70. Bd1 {[%eval -175,16] [%emt 0:00:08]} Rh4 {[%eval -252,13] [%emt 0:00:00]}
71. Bxh6 {[%eval -225,16] [%emt 0:00:04]} g4 {
(Ke5) [%eval -267,16] [%emt 0:00:07]} 72. Bf4 {[%eval -224,14] [%emt 0:00:07]}
Kf5 {[%eval -267,14] [%emt 0:00:00]} 73. Bc7 {
(Bd6) [%eval -260,14] [%emt 0:00:03]} Rxh5 {
(Kg5) [%eval -270,15] [%emt 0:00:02]} 74. Kc3 {[%eval -249,14] [%emt 0:00:03]}
Rh1 {(Kg5) [%eval -270,17] [%emt 0:00:05]} 75. Bc2+ {
(Kd2) [%eval -263,13] [%emt 0:00:04]} Ke6 {(Kg5) [%eval -310,16] [%emt 0:00:03]
} 76. Kd2 {[%eval -275,13] [%emt 0:00:03]} Rg1 {[%eval -310,17] [%emt 0:00:03]}
77. Bb6 {(Ke3) [%eval -292,14] [%emt 0:00:04]} Rg2+ {
[%eval -378,15] [%emt 0:00:03]} 78. Kd3 {(Kc3) [%eval -323,15] [%emt 0:00:07]}
g3 {(Rg3+) [%eval -407,18] [%emt 0:00:04]} 79. Bd1 {
(Be3) [%eval -332,15] [%emt 0:00:05]} Ra2 {[%eval -407,20] [%emt 0:00:04]} 80.
Bf3 {[%eval -331,15] [%emt 0:00:05]} g2 {[%eval -447,20] [%emt 0:00:10]} 81.
Bc5 {(Bd4) [%eval -418,15] [%emt 0:00:06]} f5 {[%eval -478,20] [%emt 0:00:05]}
82. Bd4 {[%eval -412,15] [%emt 0:00:03]} f4 {
(Kd6) [%eval -486,20] [%emt 0:00:03]} 83. Bc5 {[%eval -446,16] [%emt 0:00:03]}
Ra3+ {(Ke5) [%eval -486,18] [%emt 0:00:00]} 84. Ke2 {
[%eval -446,17] [%emt 0:00:05]} Ra5 {[%eval -486,20] [%emt 0:00:04]} 85. Bd4 {
[%eval -460,16] [%emt 0:00:04]} Ra2+ {[%eval -486,18] [%emt 0:00:00]} 86. Kd3 {
[%eval -460,16] [%emt 0:00:00]} Kd6 {[%eval -512,19] [%emt 0:00:04]} 87. Bb6 {
[%eval -460,15] [%emt 0:00:00]} Ra3+ {(Ra6) [%eval -512,17] [%emt 0:00:03]} 88.
Ke2 {[%eval -497,17] [%emt 0:00:05]} Ra6 {[%eval -512,18] [%emt 0:00:11]} 89.
Bd4 {[%eval -545,16] [%emt 0:00:05]} Ra2+ {[%eval -512,16] [%emt 0:00:07]} 90.
Kd3 {[%eval -545,16] [%emt 0:00:00]} Kc6 {[%eval -512,15] [%emt 0:00:04]} 91.
Bg1 {(Bh5) [%eval -554,16] [%emt 0:00:04]} Ra1 {[%eval -548,16] [%emt 0:00:03]}
92. Bd4 {[%eval -554,16] [%emt 0:00:00]} g1=Q {[%eval -593,17] [%emt 0:00:07]}
93. Bxg1 {[%eval -554,15] [%emt 0:00:00]} Rxg1 {[%eval -551,16] [%emt 0:00:02]}
94. Kd4 {[%eval -554,15] [%emt 0:00:00]} Kd6 {
(Rg5) [%eval -593,15] [%emt 0:00:03]} 95. Be2 {
(Bh5) [%eval -32752,19] [%emt 0:00:13]} Rb1 {[%eval -607,14] [%emt 0:00:09]}
96. Bg4 {[%eval -913,19] [%emt 0:00:07]} Rb4+ {
(Ra1) [%eval -615,13] [%emt 0:00:07]} 97. Kc3 {
(Kd3) [%eval -991,18] [%emt 0:00:09]} Re4 {[%eval -681,13] [%emt 0:00:06]} 0-1

Code: Select all

Just add [D] in the begining of the FEN line....

[George Tsavdaris]

In Blitz time controls on single core there was nothing else to be expected.

•Why it matters if it was in blitz time controls?
Do you suggest Rybka 3 is good at blitz while at long time controls it isn't?
Or that Crafty at blitz is bad but at longer time controls is better?
If not any of the above, then what it matters with what time control they played?

•Also why it matters if it was on single CPU?
Do you suggest Rybka 3 is good at single CPUs while at multiple CPUs/Cores it isn't?
Or that Crafty at a single CPU is bad but at multiple CPUs is better?
If not any of the above, then what it matters on how many CPUs they played?

Rybka 2 would most probably achieve a similar result.

In the exact same positions with that match i have running now a current score Rybka 3 - Rybka 2.2n2 +3 =2 -0 in favor of Rybka 3.
Which means a score of Rybka 3 - Rybka 2.2n2 4-1 for an impressive 80%.
With these i'm not trying to prove or disprove the above quote.

[Ovyron]

The new Rybka style is quite aggrssive, much more fun to watch.

Rybka human is my favorite

[Dr.Ex]

In Blitz time controls on single core there was nothing else to be expected.

•Why it matters if it was in blitz time controls?
Do you suggest Rybka 3 is good at blitz while at long time controls it isn't?
Or that Crafty at blitz is bad but at longer time controls is better?
If not any of the above, then what it matters with what time control they played?

It is certainly optimized for Bullet and Blitz play.
I bet it is not nearly that much ahead of the competition (which is certainly not Crafty) at tournament time controls.
It matters a lot at what time control engines play.

Or that Crafty at a single CPU is bad but at multiple CPUs is better?
If not any of the above, then what it matters on how many CPUs they played?

Crafty is much better on multiple CPUs but certainly still has not a chance against Rybka.

[George Tsavdaris]

In Blitz time controls on single core there was nothing else to be expected.

•Why it matters if it was in blitz time controls?
Do you suggest Rybka 3 is good at blitz while at long time controls it isn't?
Or that Crafty at blitz is bad but at longer time controls is better?
If not any of the above, then what it matters with what time control they played?

It is certainly optimized for Bullet and Blitz play.
I bet it is not nearly that much ahead of the competition (which is certainly not Crafty) at tournament time controls.

I disagree with the above you've said.
We just have NO DATA for that.
I'm not saying what you are saying is wrong or not. I'm saying that we just don't know if it's true or wrong.

But judging from Rybka 2.3.2a on CEGT and CCRL lists on blitz time controls and on longer time controls and if Rybka's 3 behavior remains the same(i see no reason for the opposite and in fact the much improved knowledge is a good reason to expect that) what you say is wrong!
Rybka seems even to gain in CEGT 40/120 list with classic time controls, more advantage from its top opponents than in blitz.

Crafty is much better on multiple CPUs but certainly still has not a chance against Rybka.

What is the evidence that Crafty is (relatively)much better on multiple CPUs than single CPU?

[Dr.Ex]

[D] 5rk1/ppp3bp/1n1q2p1/3P1n2/2P5/3BB1pP/PP1Q1P2/R3K2R w KQ - 0 19

This is a typical Crafty game btw. Crafty has a horrible evaluation here.
It castles short in this position and even thinks white is better while it is totally lost.

The line 19.0-0-0 Nxe3 20.Qxe3 Rxf2 is not nearly as bad as the game continuation.

[bob]

In Blitz time controls on single core there was nothing else to be expected.

•Why it matters if it was in blitz time controls?
Do you suggest Rybka 3 is good at blitz while at long time controls it isn't?
Or that Crafty at blitz is bad but at longer time controls is better?
If not any of the above, then what it matters with what time control they played?

It is certainly optimized for Bullet and Blitz play.
I bet it is not nearly that much ahead of the competition (which is certainly not Crafty) at tournament time controls.
It matters a lot at what time control engines play.

Or that Crafty at a single CPU is bad but at multiple CPUs is better?
If not any of the above, then what it matters on how many CPUs they played?

Crafty is much better on multiple CPUs but certainly still has not a chance against Rybka.

Crafty _always_ "has a chance".

[Graham Banks]

Crafty _always_ "has a chance".

That's the fighting spirit Bob!