Not logged inRybka Chess Community Forum
Forum RybkaChess Help Search Register Login
Up Topic The Rybka Lounge / Correspondence Chess / Relay Chess: H1N1 Junkies Vs. NATIONAL12 & Vempele 1/2-1/2
1 2 3 4 5 6 7 Previous Next  
Parent - By NATIONAL12 (Gold) [gb] Date 2009-05-15 23:24
just a general observation,when we all played chess on Hiarcs forum,i often noticed a big change in evals between d21 & d22 this has not happened in these 2 games.
Parent - - By Uly (Gold) [mx] Date 2009-05-17 04:03
Oh well, I hope Bobby doesn't kill me for this, but I have to play this move. A non-Rybka move, because Rybka's plan seems to be to draw the game or something, and I think this is the only move that keeps white's winning chances.

After this move, Rybka swims in a sea of 0.00's that are not real if you play the right moves, now, I suspect it's the position in where PCS or me alone would perform better than relayed (I think I'm not following his plan, and if he doesn't follow mine the 0.00's will become real.)

Without further ado:

22.Rd1!?

1. e4 g6 2. d4 d6 3. Nc3 Bg7 4. f4 Nf6 5. Nf3 c5 6. Bb5+
Bd7 7. e5 Ng4 8. Bxd7+ Qxd7 9. h3 cxd4 10. Qxd4 Nc6 11. Qa4
Nh6 12. Bd2 Nf5 13. O-O-O dxe5 14. Rhe1 Nfd4 15. Be3 Rd8
16. fxe5 b5 17. Qa6 Nb4 18. Qa3 Nbxc2 19. Rxd4 Nxa3
20. Rxd7 Rxd7 21. bxa3 O-O 22. Rd1

5rk1/p2rppbp/6p1/1p2P3/8/P1N1BN1P/P5P1/2KR4 b - - 0 22
Parent - - By Bobby C (****) Date 2009-05-17 04:47 Edited 2009-05-17 05:18
Perfect, and this does not result in a sea of 0.00's just yet. Black still has to play very solid and beyond the point that I have searched to get a 0.00 score. This works great for me and the move I wanted. On my machine both Rybka and Naum both like this move, I am not sure if they would change their mind at different depths or with different settings or tablebases.
Parent - - By Uly (Gold) [mx] Date 2009-05-17 05:55
Phew, that means we're on the same page, which is good, the other moves were going steadily towards 0, and I didn't see any progress, while this one stayed 0.00 around here, but upon examination the score jumps in white's favor eventually, and white does actually seem to be doing something.

I was just worried because in my hardware Rd1 is not in the top 6 moves from Rybka 3 after interactive analysis, but it's relieving that she finds it on faster hardware (so it must be the best move.)
Parent - By Vempele (Silver) [fi] Date 2009-05-17 06:01

> I was just worried because in my hardware Rd1 is not in the top 6 moves from Rybka 3 after interactive analysis


Weird, it was in the top 2 here.
Parent - By Bobby C (****) Date 2009-05-17 06:11
Now to get past the next couple moves on the other board. :)
Parent - By NATIONAL12 (Gold) [gb] Date 2009-05-17 07:18
22...Rxd1+
Parent - - By Bobby C (****) Date 2009-05-17 08:11
23.Kxd1

5rk1/p3ppbp/6p1/1p2P3/8/P1N1BN1P/P5P1/3K4 b - - 0 23


FYI, for those who are interested. Here is the speed of my system in this position with Naum 4 at depth 28. I deleted all analysis and evaluations because of the rules of the game. Also if you post this kind information please do the same. I will post the same speed information when I get a chance with Rybka.

Naum:

Depth      Time         Total Nodes               Nodes/Sec 
28/46     08:42      3.507.732,952             6.714.067 
Parent - By Vempele (Silver) [fi] Date 2009-05-17 08:50 Edited 2009-05-17 09:13
23...a6

Wow, Rybka really sucks in this endgame. Nonzero KBNPkrp.
Parent - - By Uly (Gold) [mx] Date 2009-05-18 12:37
Ok, now this is exciting, I think we could win this one!



24.a4

1. e4 g6 2. d4 d6 3. Nc3 Bg7 4. f4 Nf6 5. Nf3 c5 6. Bb5+
Bd7 7. e5 Ng4 8. Bxd7+ Qxd7 9. h3 cxd4 10. Qxd4 Nc6 11. Qa4
Nh6 12. Bd2 Nf5 13. O-O-O dxe5 14. Rhe1 Nfd4 15. Be3 Rd8
16. fxe5 b5 17. Qa6 Nb4 18. Qa3 Nbxc2 19. Rxd4 Nxa3
20. Rxd7 Rxd7 21. bxa3 O-O 22. Rd1 Rxd1+ 23. Kxd1 a6 24. a4

5rk1/4ppbp/p5p1/1p2P3/P7/2N1BN1P/P5P1/3K4 b - -
Parent - - By NATIONAL12 (Gold) [gb] Date 2009-05-18 12:59
in your dreams24...bxa4
Parent - By Uly (Gold) [mx] Date 2009-05-18 14:43
Yeah well, it would be really boring to play without any hopes, I still think that was the main reason the Hiarcs forum games ended in draws, people were expecting draws and that's what they got (not counting games won on time), here I'm trying to win, and I hope Bobby will do too.
Parent - - By Bobby C (****) Date 2009-05-19 06:30
25.Nxa4 

1. e4 g6 2. d4 d6 3. Nc3 Bg7 4. f4 Nf6 5. Nf3 c5 6. Bb5+
Bd7 7. e5 Ng4 8. Bxd7+ Qxd7 9. h3 cxd4 10. Qxd4 Nc6 11. Qa4
Nh6 12. Bd2 Nf5 13. O-O-O dxe5 14. Rhe1 Nfd4 15. Be3 Rd8
16. fxe5 b5 17. Qa6 Nb4 18. Qa3 Nbxc2 19. Rxd4 Nxa3
20. Rxd7 Rxd7 21. bxa3 O-O 22. Rd1 Rxd1+ 23. Kxd1 a6 24. a4 bxa5
25.Nxa4

5rk1/4ppbp/p5p1/4P3/N7/4BN1P/P5P1/3K4 b - - 0 25
Parent - - By Vempele (Silver) [fi] Date 2009-05-19 13:06
25...Rc8 0.00
Parent - - By Bobby C (****) Date 2009-05-19 13:13
0.00 with a hell of a lot of forward pruning :) But, probably a draw, let's see :)
Parent - - By Vempele (Silver) [fi] Date 2009-05-19 13:16
0.00 with 7000 posions in IDeA, no significant human guidance other than setting a few variations to 0.00 because Rybka doesn't understand insufficient material [advantage] in pawnless endgames.
Parent - - By Bobby C (****) Date 2009-05-19 13:20
Yes, Rybka does not have much of a clue. If Naum or Sjeng does not come up with something fast..... :)
Parent - - By Vempele (Silver) [fi] Date 2009-05-19 13:53
Gah, Deep Prolongation with 100 positions in the queue. Must be some kind of record.
Parent - - By chesscomacc (*) [in] Date 2009-05-19 14:27 Edited 2009-05-19 14:30
Wow, white WON!!!... according to game played with Zappa Mexico 2 x64 bit!! white has advantage of 5.30 point over black, game played with zappa itself above depth 19 on core i7 920 .. here is the move list.

25. ... Rc8 26. Kd2 e6 27. Kd3 Bf8 28. Nb6 Rc7 29. Nc4 Rd7+ 30. Bd4 Bb4 31.
a3 Bc5 32. Kc3 Bxd4+ 33. Nxd4 h5 34. a4 g5 35. Nf3 g4 36. hxg4 hxg4 37. Nh2
g3 38. Nf1 Rd1 39. Nxg3 Re1 40. Kd4 Rd1+ 41. Kc5 Rb1 42. Ne4 Kf8 43. a5 Ke7
44. Ncd6 Rb3 45. Kc6 Rb1 46. g4 Rg1 47. Kb7 Rxg4 48. Kxa6 *

ultimately leading to passed pawn, black will have to sacrifice rook now.
here is the last analysis and position of the game.

8/4kp2/K2Np3/P3P3/4N1r1/8/8/8 b - - 0 48


Zappa Mexico II x64:
  19  00:00            19  0  -4.05  Rg4g2 Ka6b6 Rg2b2+ Kb6a7 Rb2b4 Ne4f6 Ke7f8 Nf6h7+ Kf8e7 a5a6 Rb4b1 Nh7f6 Ke7f8 Nf6e4 Kf8e7 Ka7a8 Ke7d7 a6a7 Kd7e7 Ne4f6
20/45  00:07    36.230.929  5.085.043  -4.73  Rg4g2 Ka6b6 Rg2b2+ Nd6b5 Rb2a2 a5a6 Ke7f8 a6a7 Kf8g7 Ne4c5 Ra2xa7 Nb5xa7 Kg7g6 Kb6c6 Kg6f5 Kc6d6 Kf5g4 Nc5xe6
20/45  00:19    93.768.399  4.808.636  -4.73  Rg4g2 Ka6b6 Rg2b2+ Nd6b5 Rb2a2 a5a6 Ke7f8 a6a7 Kf8g7 Ne4c5 Ra2xa7 Nb5xa7 Kg7g6 Kb6c6 Kg6f5 Kc6d6 Kf5g4 Nc5xe6 Kg4h5
21/52  00:41   207.788.957  4.980.799  -5.10  Rg4g2 Ka6b6 Rg2b2+ Nd6b5 Rb2a2 a5a6 Ke7f8 a6a7 Kf8g7 Ne4c5 Ra2xa7 Nb5xa7 Kg7g6 Kb6c7 f7f6 Kc7d6 Kg6f5 e5xf6 Kf5xf6 Nc5xe6 Kf6f5 Kd6d5 Kf5g4 Na7c6
21/52  00:56   277.234.937  4.908.203  -5.10  Rg4g2 Ka6b6 Rg2b2+ Nd6b5 Rb2a2 a5a6 Ke7f8 a6a7 Kf8g7 Ne4c5 Ra2xa7 Nb5xa7 Kg7g6 Kb6c7 f7f6 Kc7d6 Kg6f5 e5xf6 Kf5xf6 Nc5xe6 Kf6f5 Kd6d5 Kf5g4 Na7c6
22/52  01:02   306.300.696  4.919.308  -5.10  Rg4g2 Ka6b6 Rg2b2+ Nd6b5 Rb2a2 a5a6 Ke7f8 a6a7 Kf8g7 Ne4c5 Ra2xa7 Nb5xa7 Kg7g6 Kb6c7 f7f6 Kc7d6 Kg6f5 e5xf6 Kf5xf6 Nc5xe6 Kf6f5 Kd6d5 Kf5g4 Na7c6
22/52  01:18   367.420.611  4.680.517  -5.10  Rg4g2 Ka6b6 Rg2b2+ Nd6b5 Rb2a2 a5a6 Ke7f8 a6a7 Kf8g7 Ne4c5 Ra2xa7 Nb5xa7 Kg7g6 Kb6c7 f7f6 Kc7d6 Kg6f5 e5xf6 Kf5xf6 Nc5xe6 Kf6f5 Kd6d5 Kf5g4 Na7c6
23/52  01:38   460.910.008  4.672.648  -5.10  Rg4g2 Ka6b6 Rg2b2+ Nd6b5 Rb2a2 a5a6 Ke7f8 a6a7 Kf8g7 Ne4c5 Ra2xa7 Nb5xa7 Kg7g6 Kb6c7 f7f6 Kc7d6 Kg6f5 e5xf6 Kf5xf6 Nc5xe6 Kf6f5 Kd6d5 Kf5g4 Na7c6
23/52  02:07   578.836.104  4.542.670  -5.10  Rg4g2 Ka6b6 Rg2b2+ Nd6b5 Rb2a2 a5a6 Ke7f8 a6a7 Kf8g7 Ne4c5 Ra2xa7 Nb5xa7 Kg7g6 Kb6c7 f7f6 Kc7d6 Kg6f5 e5xf6 Kf5xf6 Nc5xe6 Kf6f5 Kd6d5 Kf5g4 Na7c6
24/58  04:12  1.237.950.881  4.899.147  -5.30  Rg4g2 Ka6b6 Rg2b2+ Nd6b5 Rb2a2 a5a6 Ke7f8 a6a7 Kf8g7 Ne4c5 Ra2xa7 Nb5xa7 f7f6 Nc5xe6+ Kg7h6 Na7c6 f6xe5 Nc6xe5 Kh6h5 Kb6c5 Kh5h4 Kc5d4 Kh4g3

(Note : Zappa displays scores accoriding to the person whose move is, currently black has to move... and have score -5.30 at depth 24/58)
Parent - By chesscomacc (*) [in] Date 2009-05-19 14:55
Rybka 3 first thinks of Rb8, then changes to f7-f6, and then rd8+, If black plays Rd8= then Zappa changed to Ke2 !!

This is astonishing, !! Zappa already is thinking of Rc8 from the first move at depth 14 !! and already a winning score to white!! whats this?? Zappa is so powerful in endgame? !!
Parent - By chesscomacc (*) [in] Date 2009-05-19 15:16 Edited 2009-05-19 15:33
Ok everything clarified, Rybka is not showing DRAW!!! It is showing a very less value, that doesnt mean its adraw,

Rybka scoring is different from other engines, when Rybka play with itself, white won!

A good thing to point about, when Zappa exchanged bishops earlier (after 25) but Rybka kept 1 bishop and 1 knight, Zappa sacrificed the rook earlier to a passed pawn, Rybka did'nt do that.

Vempele, you might regret your words about Rybka and its endgame, :)
Parent - - By Vempele (Silver) [fi] Date 2009-05-19 15:45
Heh, your game deviates from my analysis before move 30 (might be 26, might be 29, might be somewhere in between. Not saying until the move (and perhaps a few more) is on the board.
Parent - - By chesscomacc (*) [in] Date 2009-05-19 15:48
Yea, there are lots of variations possible here, and most are equally somewhat 'Same' ....

Like 25....Rd8+ is also fine.
Parent - By Vempele (Silver) [fi] Date 2009-05-19 16:07
No, the one I'm talking about is an only move.
Parent - - By NATIONAL12 (Gold) [gb] Date 2009-05-19 17:33
i wont be playing Zappas line.
Parent - By chesscomacc (*) [in] Date 2009-05-19 18:11
At this position, white is ahead in mobility... all opened position for Bishop and knights. I think Rybka ignores the mobility factor a little bit in endgame... thats why is showing less score. anyway black should open its bishop and rook...f6 ?
Parent - - By Uly (Gold) [mx] Date 2009-05-20 00:38

> 0.00


Still in the sea of 0.00's? :twisted:

26.Kd2 0.18

1. e4 g6 2. d4 d6 3. Nc3 Bg7 4. f4 Nf6 5. Nf3 c5 6. Bb5+
Bd7 7. e5 Ng4 8. Bxd7+ Qxd7 9. h3 cxd4 10. Qxd4 Nc6 11. Qa4
Nh6 12. Bd2 Nf5 13. O-O-O dxe5 14. Rhe1 Nfd4 15. Be3 Rd8
16. fxe5 b5 17. Qa6 Nb4 18. Qa3 Nbxc2 19. Rxd4 Nxa3
20. Rxd7 Rxd7 21. bxa3 O-O 22. Rd1 Rxd1+ 23. Kxd1 a6 24. a4
bxa4 25. Nxa4 Rc8 26. Kd2

2r3k1/4ppbp/p5p1/4P3/N7/4BN1P/P2K2P1/8 b - -
Parent - - By Vempele (Silver) [fi] Date 2009-05-20 01:24

> Still in the sea of 0.00's? :twisted:


There's no sea anymore. I created a set of simple axioms from which I've derived a drawing strategy.

Took me about 5 minutes to confirm the draw once I had the axioms. :evil:
Parent - By Uly (Gold) [mx] Date 2009-05-20 02:21
But you're not allowed to tell the axioms to NATIONAL12 :razz:
Parent - By NATIONAL12 (Gold) [gb] Date 2009-05-20 06:25
26...f6
Parent - By Bobby C (****) Date 2009-05-20 13:23

> Still in the sea of 0.00's?


Maybe your using the wrong chess engine. :) I can not find a line that scores 0.00 with any engine I am using, including Rybka.
Parent - By keoki010 (Bronze) [us] Date 2009-05-19 15:26
PCS you've got a mistake on 24...bxa5 it should be 24...bxa4
Parent - - By Bobby C (****) Date 2009-05-20 13:19
27.e6
Parent - - By chesscomacc (*) [in] Date 2009-05-20 13:26

> Maybe your using the wrong chess engine. :-) I can not find a line that scores 0.00 with any engine I am using, including Rybka.


There is one line, it is 26 ... f6 27. e6 , Then its a draw!
Parent - - By Bobby C (****) Date 2009-05-20 20:40
I can promise that no chess engine I am using scores this position as 0.00, or less then 0.10 at higher then depth 30 after some forward pruning. Now this position might be a draw but I will not agree to that until all of the top chess engines are seeing this. Two chess engines that are in the top five in the world have this at > 0.25 and Rybka itself has it > 0.10. Vempele has some good OTB endgame skills so if he thinks the position was a forced draw by black several moves ago then I can not tell him otherwise, just that a few 3000+ elo chess engines disagree with him. :)
Parent - By NATIONAL12 (Gold) [gb] Date 2009-05-20 20:53
i agree with you,R3 at d32 gives 0.10 and Naums eval are a bit higher
Parent - - By chesscomacc (*) [in] Date 2009-05-21 03:31 Edited 2009-05-21 04:30

> I can promise that no chess engine I am using scores this position as 0.00, or less then 0.10 at higher then depth 30 after some forward pruning. Now this position might be a draw but I will not agree to that until all of the top chess engines are seeing this. Two chess engines that are in the top five in the world have this at > 0.25 and Rybka itself has it > 0.10. Vempele has some good OTB endgame skills so if he thinks the position was a forced draw by black several moves ago then I can not tell him otherwise, just that a few 3000+ elo chess engines disagree with him. :-)


Yea, you are right, Zappa 2 is reporting score of +1.1 !!, but Rybka 3 is showing 0.00, whats this difference about? She has found a draw line that black can take advantage of.

Rybka 3 x64 8 CPU is giving a draw score here 0.00 here is the analysis

FEN:
2r3k1/4p1bp/p3Ppp1/8/N7/4BN1P/P2K2P1/8 b - - 0 1


Rybka3 x64:

[Deleted /]
Parent - - By Uly (Gold) [mx] Date 2009-05-21 04:14
Note that I'm not looking at these posted analysis and I would expect other players to do the same (not being able to comunicate should include outsiders.)
Parent - - By Bobby C (****) Date 2009-05-21 04:24 Edited 2009-05-21 04:30
Edit his post and delete this stuff. Also warn him about outside help, apparently he did not read the intro to the games. BTW, the stuff he has posted is basically useless to me or anyone else so it has not hurt anything. But to avoid possible issues in the future let's make sure he knows not to do it again....
Parent - - By chesscomacc (*) [in] Date 2009-05-21 10:22
I am sorry, I wont post analysis when game is going on, but that analysis was just a draw line shown by Rybka....I wanted to ask why difference b/w zappa, naum and Rybka..
Parent - By Bobby C (****) Date 2009-05-21 11:12

> I wanted to ask why difference b/w zappa, naum and Rybka..


There is a difference because they are different engines. Some search faster and some play certain positions better. Rybka is sometimes terrible in endgames but it searches so much better and so much faster then the others.
Parent - - By NATIONAL12 (Gold) [gb] Date 2009-05-20 22:38
it must have been a difficult choice between e6 and Kd3.
Parent - - By Bobby C (****) Date 2009-05-21 10:22
Not really, there are 3 engines in the top 5 that can play this endgame better then Rybka , IMHO. All three preferred e6 over Kd3, but maybe this can lead to a type of transposition. One thing I would like to do after the game if it ends in a draw is to find someone with equal or better hardware to play me a match against the other three engines using only Rybka. I think one of those engines will win and I have not got two quads laying around to do the test myself. 2CPU vs 2CPU would not be reliable since at 120min per game would not be enough for the engines to get past depth 21 or so per move.
Parent - - By NATIONAL12 (Gold) [gb] Date 2009-05-21 21:21
i might be able to take you up on this,the problem is the opening books.if we could reach an agreement on a position to start with it could be an intresting game.i would be happy to use Rybka playing black.
Parent - - By NATIONAL12 (Gold) [gb] Date 2009-05-21 23:16
Vas & Lukas(Kullberg) believe they can win freestyle against anything with Lukas cluster,i am fairly sure you cannot beat me with other engines.i will use  Rybka with some tbs.could create a lot of intrest in the forum.one of your moves has already been criticized by an OTB GM.great fun to me.i know how much you would like to beat me.
Parent - - By Bobby C (****) Date 2009-05-22 09:58

> one of your moves has already been criticized by an OTB GM


Where can I find this at? What move are we talking about?
Parent - By Uly (Gold) [mx] Date 2009-05-22 14:53 Edited 2009-05-22 15:21

> Where can I find this at?


[Deleted - Sent by PM]
Parent - By Bobby C (****) Date 2009-05-22 10:00
The starting position is the position this game is currently at. I was not challenging anyone to a Engine x vs Rybka game. I can do that anytime on playchess. I was challenging an Engine x vs Rybka game using this endgame as the start position, basically seeing if Rybka can play this endgame better then other engines.
Parent - - By Vempele (Silver) [fi] Date 2009-05-21 05:48
27...Rc4 0.00
Parent - - By AsosLight (***) [gr] Date 2009-05-21 12:58
To me 27.Kd3 was essential, with a long way to go.
Up Topic The Rybka Lounge / Correspondence Chess / Relay Chess: H1N1 Junkies Vs. NATIONAL12 & Vempele 1/2-1/2
1 2 3 4 5 6 7 Previous Next  

Powered by mwForum 2.27.4 © 1999-2012 Markus Wichitill