Not logged inRybka Chess Community Forum
Up Topic Rybka Support & Discussion / Rybka Discussion / Rybka @ Leiden ICT7
- - By Vasik Rajlich (Silver) [hu] Date 2007-05-20 19:22
The 7th International CSVN Tournament has finished and Rybka has won with 7.5/9, ahead of Zappa, Sjeng and Hiarcs.

The full Rybka team was:

Operator: Hans van der Zijden
Opening book: Jeroen Noomen
Engine help: Larry Kaufman
Engine & Hardware: Vasik Rajlich

I'd like to thank:

- Theo van Storm for organizing the event and keeping everything running smoothly
- The other teams who took the time to participate
- Bjorn Osterman for offering the use of his 16-way Opteron (unfortunately, we were not able to solve the technical problems)
- The members of the Rybka team for their hard work

The full tournament crosstable can be found here:

http://www.csvn.nl/index.php?option=com_content&task=view&id=258&Itemid=28

Here is a quick roundup of Rybka's games:

Rd 1, The Baron - Rybka 1/2-1/2: Black equalized easily in a theoretical Gruenfeld line and gradually turned the quiet symmetrical-pawns position in her favor. Then, in a winning rook endgame, Mr. Murphy came for his first visit. Due to problems with the remote Rybka 2.3.2 Beta version, we switched the seemingly routine mop-up duty to our on-site machine, where Rybka blundered and allowed black to escape to a tablebase draw.

Rd 2, Rybka - Micro-Max 1-0: Micro-Max is allegedly the most compact chess program in the world. I'm not sure if that precludes an opening book, anyway black's opening variation looks suspicious and his position collapsed quickly.

Rd 3, Hiarcs - Rybka 1/2-1/2: The Hiarcs team apparently decided not to mess with Jeroen this round. A quiet main-line 2. c3 Sicilian led to a dead-drawn QN vs QB endgame.

Rd 4, Rybka - Stoofvlees 1-0: Stoofvlees seems to be a brainstorm of Sjeng author Gian-Carlo Pascutto, making its debut here. Black grabbed a very risky pawn in an open Sicilian and white's attack crashed through.

Rd 5, Zappa - Rybka 0-1: Erdo also apparently decided not to mess with Jeroen this round :), opting for 5. h3 in a Closed Sicilian. White's opening setup turned out to be a bit awkward. Zappa pressed forward but neglected his queenside pieces and black's counterattack crashed through. Black's queenside bishop did not move while black's queenside rook moved only to a3 until after the game was decided.

Rd 6, Rybka - Sjeng 1/2-1/2: The opening was a rout for white, but things went downhill quickly. First, Rybka pushed her queenside pawn too quickly, and black could have drawn with 31. .. Nd7! 32. Ba6 Rxb6 (32. .. Qa8 33. b7 Qa7+ 34. Kh1 Qd4 should also be good enough to draw) 33. Bxb7 Rxb1+ 34. Kg2 Rxb7 and white won't win in a million years. Sjeng also overlooked this, setting up a finish to the game which could have come out of a computer chess book from the 1970s. Due to a bug, Rybka locked in to a score of +1.75 and moved around aimlessly, missing several quick finishes. The last really clear win slipped with 63. Kc5, and after 66. Qd6 black no longer had to worry.

Rd 7, Rybka - IsiChess 1-0: Black's opening novelty 12. .. Nd8 appears to leave his position in ruins. Black tried to wiggle out into a two pieces vs rook and pawn endgame, but white's pieces were too active.

Rd 8, Glaurung - Rybka 0-1: After a tame symmetrical queen pawn opening white uncorked the wild shot 21. Rxe6. With an open e-file on the board, the idea looks a bit dubious, but in fact several moves later Jeroen claims that Glaurung could have forced a complex draw by playing 25. Nxd5. Instead, Glaurung sacrificed his queen and never seemed to have enough compensation.

Rd 9, Rybka - Scorpio 1-0: Scorpio went for 10. .. a5 in the Najdorf. I don't like the look of black's position at all, but maybe a Najdorf specialist could correct me on this. White has a simple plan of blockading the light squares, and this is exactly what happened in the game. Black could not free himself and was ground down.

Vas
Parent - By Dr.X (Gold) Date 2007-05-20 19:29 Edited 2007-05-20 19:33
Vas,

Terrific! I am proud to be a Rybka-ian. Congratulations on yet another pièce de résistance served up by Rybka.

sidserious
Parent - - By exigentsky (***) [us] Date 2007-05-20 19:33
Congratulations Vas and team! Rybka has proven its worth once again. :)

But I'm curious, will the improved version of Rybka's book that Jeroen mentioned perhaps be released with Rybka 2.3.2?

BTW: What was the time control at Leiden?
Parent - - By sam_i_am (**) Date 2007-05-20 19:53
You mean that the Fritz & Junior team did not participate?

Damned shame you know...
Parent - - By Dr.X (Gold) Date 2007-05-21 02:02
Sam,

Where have you been, man! Fritz & Junior if they did participate they would have to concede  Rybka's superiority.

But, I think the the mumblings of some that  Zap!Chess Zanzibar, although very good,  is not quite good enough to beat Rybka.

I wonder what happened also to Spike?

sidserious
Parent - - By sam_i_am (**) Date 2007-05-21 05:15 Edited 2007-05-21 05:19
Would also like to know what has happened to the development of Spike.

There isn't any news on the spike page and e-mail after e-mail remains unanswered.(I'm sure I'm not the only one )

Maybe Volker Boehm & Ralf Schaefer will just sit on their private versions until it becomes obsolete.
Parent - By Dr.X (Gold) Date 2007-05-21 17:23
Sam,

I think you are right on the money with that. I have sent them numerous email without one reply. I'll bet they were very upset with Ridderkerk stating that they were the strongest "Free" engine. No doubt they are distancing themselves from that "Free" idea to one of $$$$$$$$$$$.

Good to hear that you are still with us.

sidserious
Parent - - By exigentsky (***) [us] Date 2007-05-21 18:21
Come on Vas, can we please get Jeroen's improved book? :P
Parent - - By Vasik Rajlich (Silver) [hu] Date 2007-05-23 07:12
This is of course up to Jeroen. I haven't discussed it with him yet. Probably, RybkaIII.ctg will be released with Rybka3.exe, but please don't quote me on this.

Vas
Parent - - By DamirD81 (***) [dk] Date 2007-05-23 09:06
Hi Vas

Maybe you should just name it Rybka 2.2.3 ctg instead of 3, since it is not out yet, and wait till Rybka 3 is ready, and than simply name it Rybka III ctg.
Although I doubt it will be much of a help since there already is a book called Rybka III. Maybe you should consider in calling it something else.
Parent - By exigentsky (***) [us] Date 2007-05-24 07:38
It would be great if we could get the updated book! I for one would be extremely pleased. Rybka III is still a long time in the future. Please just release it as something like Rybka 2.3.2 or RYBKA_II_UPDATED. But please do release it. What do you say Jeroen?
Parent - - By Werner (***) [de] Date 2007-05-20 19:57
Hi Vas,
first congrats to the win of ICT 7.
Is it true, you switched to version 2.2n2 during the match? Maybe we have to wait some more weeks for a new version??

regards
Werner
Parent - - By Uri Blass (*****) [il] Date 2007-05-20 23:40
I am also interested to know reply to that question.
If the target was to use version without bugs then 2.2n2 is no solution because you have hash bugs in it and you know it.

see the following thread when you thanked me for telling you about the bug
in
3k4/4r3/8/3Q4/8/1P6/1PP1p1PP/7K b - - 0 1
.

http://chessobserver.com/rybkaforum/viewtopic.php?t=562

Uri
Parent - By Gaмßito (****) [cr] Date 2007-05-21 08:17 Edited 2007-05-21 08:20
Hi,

Yes, 2.2n2 has some hash bugs. In this position it caused not to see the draw line. Old 1.2f version does not have any problem with this position, neither 1.2n, and i suppose the bug was produced when Vas add some new modifications in 2x versions. Only Vas can give us an appropriate answer to this.

Analysis by Rybka 1.2f 32-bit:

1...Kc7
  +-  (6.24)   Depth: 2   00:00:00
1...Kc7
  +-  (6.21)   Depth: 3   00:00:00
1...Kc7 2.Qc5+ Kd7
  +-  (7.63)   Depth: 4   00:00:00
1...Kc7 2.Qc5+ Kd7 3.Qb5+ Kd6
  +-  (7.63)   Depth: 5   00:00:00
1...Kc7 2.Qc5+ Kd7 3.Qb5+ Kd6 4.Qd3+ Kc6
  +-  (7.63)   Depth: 6   00:00:00  3kN
1...Kc7 2.Qc5+ Kd7 3.Qb5+ Kd6 4.Qd3+ Kc6 5.Qc3+ Kd7
  +-  (7.63)   Depth: 7   00:00:00  5kN
1...Kc7 2.Qc5+ Kd7 3.Qb5+ Kd6 4.Qd3+ Kc6 5.Qc3+ Kd7 6.Qe1
  +-  (7.63)   Depth: 8   00:00:00  10kN
1...Kc7 2.Qc5+ Kd7 3.Qb5+ Kd6 4.Qd3+ Kc6 5.Qc3+ Kd7 6.Qd2+ Kc6 7.Qe1
  +-  (7.63)   Depth: 9   00:00:00  19kN
1...Kc7 2.Qc5+ Kd7 3.Qb5+ Kd6 4.Qd3+ Kc6 5.Qc3+ Kd7 6.Qe1 Re4 7.Kg1
  +-  (7.63)   Depth: 10   00:00:00  31kN
1...Kc7 2.Qc5+ Kd7 3.Qb5+ Kd6 4.Qd3+ Kc6 5.Qc3+ Kd7 6.Qd2+ Kc6 7.Qe1 Kd7 8.Kg1
  +-  (7.63)   Depth: 11   00:00:01  50kN
1...Kc7 2.Qc5+ Kd7 3.Qb5+ Kd6 4.Qd3+ Kc6 5.Qc3+ Kd7 6.Qe1 Re4 7.Kg1 Ke6 8.c4
  +-  (7.63)   Depth: 12   00:00:01  81kN
1...Ke8 2.Qh5+ Kf8 3.Qf5+ Kg7 4.Qg5+ Kf7 5.Qf4+ Ke8 6.Qa4+ Kf7 7.Qf4+ Ke8 8.Qa4+
  =  (0.00)   Depth: 12   00:00:01  92kN
1...Ke8 2.Qh5+ Kf8 3.Qf5+ Kg7 4.Qg5+ Kf7 5.Qf4+ Ke8 6.Qa4+ Kf7 7.Qf4+ Ke8 8.Qa4+
  =  (0.00)   Depth: 13   00:00:01  100kN
1...Ke8 2.Qh5+ Kf8 3.Qf5+ Kg7 4.Qg5+ Kf7 5.Qf4+ Ke8 6.Qa4+ Kf7 7.Qf4+ Ke8 8.Qa4+
  =  (0.00)   Depth: 14   00:00:02  113kN
1...Ke8 2.Qh5+ Kf8 3.Qf5+ Kg7 4.Qg5+ Kf7 5.Qf4+ Ke8 6.Qa4+ Kf7 7.Qf4+ Ke8 8.Qa4+
  =  (0.00)   Depth: 15   00:00:02  132kN
1...Ke8 2.Qh5+ Kf8 3.Qf5+ Kg7 4.Qg5+ Kf7 5.Qf4+ Ke8 6.Qa4+ Kf7 7.Qf4+ Ke8 8.Qa4+
  =  (0.00)   Depth: 16   00:00:02  161kN
1...Ke8 2.Qh5+ Kf8 3.Qf5+ Kg7 4.Qg5+ Kf7 5.Qf4+ Ke8 6.Qa4+ Kf7 7.Qf4+ Ke8 8.Qa4+
  =  (0.00)   Depth: 17   00:00:02  200kN
1...Ke8 2.Qh5+ Kf8 3.Qf5+ Kg7 4.Qg5+ Kf7 5.Qf4+ Ke8 6.Qa4+ Kf7 7.Qf4+ Ke8 8.Qa4+
  =  (0.00)   Depth: 18   00:00:03  257kN
1...Ke8 2.Qh5+ Kf8 3.Qf5+ Kg7 4.Qg5+ Kf7 5.Qf4+ Ke8 6.Qa4+ Kf7 7.Qf4+ Ke8 8.Qa4+
  =  (0.00)   Depth: 19   00:00:03  335kN
1...Ke8 2.Qh5+ Kf8 3.Qf5+ Kg7 4.Qg5+ Kf7 5.Qf4+ Ke8 6.Qa4+ Kf7 7.Qf4+ Ke8 8.Qa4+
  =  (0.00)   Depth: 20   00:00:03  444kN
1...Ke8 2.Qh5+ Kf8 3.Qf5+ Kg7 4.Qg5+ Kf7 5.Qf4+ Ke8 6.Qa4+ Kf7 7.Qf4+ Ke8 8.Qa4+
  =  (0.00)   Depth: 21   00:00:04  600kN
1...Ke8 2.Qh5+ Kf8 3.Qf5+ Kg7 4.Qg5+ Kf7 5.Qf4+ Ke8 6.Qa4+ Kf7 7.Qf4+ Ke8 8.Qa4+
  =  (0.00)   Depth: 22   00:00:04  812kN
1...Ke8 2.Qh5+ Kf8 3.Qf5+ Kg7 4.Qg5+ Kf7 5.Qf4+ Ke8 6.Qa4+ Kf7 7.Qf4+ Ke8 8.Qa4+
  =  (0.00)   Depth: 23   00:00:05  1092kN
1...Ke8 2.Qh5+ Kf8 3.Qf5+ Kg7 4.Qg5+ Kf7 5.Qf4+ Ke8 6.Qa4+ Kf7 7.Qf4+ Ke8 8.Qa4+
  =  (0.00)   Depth: 24   00:00:07  1462kN
1...Ke8 2.Qh5+ Kf8 3.Qf5+ Kg7 4.Qg5+ Kf7 5.Qf4+ Ke8 6.Qa4+ Kf7 7.Qf4+ Ke8 8.Qa4+
  =  (0.00)   Depth: 25   00:00:09  1977kN
1...Ke8 2.Qh5+ Kf8 3.Qf5+ Kg7 4.Qg5+ Kf7 5.Qf4+ Ke8 6.Qa4+ Kf7 7.Qf4+ Ke8 8.Qa4+
  =  (0.00)   Depth: 26   00:00:13  2686kN
1...Ke8 2.Qh5+ Kf8 3.Qf5+ Kg7 4.Qg5+ Kf7 5.Qf4+ Ke8 6.Qa4+ Kf7 7.Qf4+ Ke8 8.Qa4+
  =  (0.00)   Depth: 27   00:00:17  3650kN
1...Ke8 2.Qh5+ Kf8 3.Qf5+ Kg7 4.Qg5+ Kf7 5.Qf4+ Ke8 6.Qa4+ Kf7 7.Qf4+ Ke8 8.Qa4+
  =  (0.00)   Depth: 28   00:00:24  5015kN
1...Ke8 2.Qh5+ Kf8 3.Qf5+ Kg7 4.Qg5+ Kf7 5.Qf4+ Ke8 6.Qa4+ Kf7 7.Qf4+ Ke8 8.Qa4+
  =  (0.00)   Depth: 29   00:00:32  6795kN
1...Ke8 2.Qh5+ Kf8 3.Qf5+ Kg7 4.Qg5+ Kf7 5.Qf4+ Ke8 6.Qa4+ Kf7 7.Qf4+ Ke8 8.Qa4+
  =  (0.00)   Depth: 30   00:00:44  9306kN
1...Ke8 2.Qh5+ Kf8 3.Qf5+ Kg7 4.Qg5+ Kf7 5.Qf4+ Ke8 6.Qa4+ Kf7 7.Qf4+ Ke8 8.Qa4+
  =  (0.00)   Depth: 31   00:01:02  12738kN
1...Ke8 2.Qh5+ Kf8 3.Qf5+ Kg7 4.Qg5+ Kf7 5.Qf4+ Ke8 6.Qa4+ Kf7 7.Qf4+ Ke8 8.Qa4+
  =  (0.00)   Depth: 32   00:01:30  17747kN
1...Ke8 2.Qh5+ Kf8 3.Qf5+ Kg7 4.Qg5+ Kf7 5.Qf4+ Ke8 6.Qa4+ Kf7 7.Qf4+ Ke8 8.Qa4+
  =  (0.00)   Depth: 33   00:02:10  24983kN

(, AMD 21.05.2007)

Analysis by Rybka 2.2 32-bit:

1...Kc7
  +-  (6.18)   Depth: 2   00:00:00
1...Kc7
  +-  (6.18)   Depth: 3   00:00:00
1...Kc7
  +-  (7.63)   Depth: 4   00:00:00
1...Kc7 2.Qa5+ Kc6
  +-  (7.63)   Depth: 5   00:00:00
1...Kc7 2.Qa5+ Kc6 3.Qc3+ Kd7
  +-  (7.63)   Depth: 6   00:00:00  3kN
1...Kc7 2.Qa5+ Kc6 3.Qe1 Kd7
  +-  (7.63)   Depth: 7   00:00:00  6kN
1...Kc7 2.Qa5+ Kc6 3.Qc3+ Kd7 4.Qe1 Kc6
  +-  (7.63)   Depth: 8   00:00:00  11kN
1...Kc7 2.Qa5+ Kc6 3.Qe1 Kd7 4.b4 Re5
  +-  (7.63)   Depth: 9   00:00:00  17kN
1...Kc7 2.Qa5+ Kc6 3.Qc3+ Kd7 4.Qe1 Kc6 5.b4 Kb7
  +-  (7.63)   Depth: 10   00:00:00  35kN
1...Kc7 2.Qa5+ Kc6 3.Qc3+ Kd7 4.Qd3+ Kc7 5.Qc4+ Kd7 6.Qb5+ Kc7 7.Qc5+ Kd7 8.Qa7+
  +-  (7.63)   Depth: 11   00:00:01  48kN
1...Kc7 2.Qa5+ Kc6 3.Qc3+ Kd7 4.Qd3+ Kc7 5.Qc4+ Kd7 6.Qb5+ Kc7 7.Qc5+ Kd7 8.Qa7+
  +-  (7.63)   Depth: 12   00:00:01  55kN
1...Kc7 2.Qa5+ Kc6 3.Qc3+ Kd7 4.Qd3+ Kc7 5.Qc4+ Kd7 6.Qb5+ Kc7 7.Qc5+ Kd7 8.Qa7+
  +-  (7.63)   Depth: 13   00:00:01  66kN
1...Kc7 2.Qa5+ Kc6 3.Qc3+ Kd7 4.Qd3+ Kc7 5.Qc4+ Kd7 6.Qb5+ Kc7 7.Qc5+ Kd7 8.Qa7+
  +-  (7.63)   Depth: 14   00:00:02  84kN
1...Kc7 2.Qa5+ Kc6 3.Qc3+ Kd7 4.Qd3+ Kc7 5.Qc4+ Kd7 6.Qb5+ Kc7 7.Qc5+ Kd7 8.Qa7+
  +-  (7.63)   Depth: 15   00:00:02  107kN
1...Kc7 2.Qa5+ Kc6 3.Qc3+ Kd7 4.Qd3+ Kc7 5.Qc4+ Kd7 6.Qb5+ Kc7 7.Qc5+ Kd7 8.Qa7+
  +-  (7.63)   Depth: 16   00:00:02  148kN
1...Kc7 2.Qa5+ Kc6 3.Qc3+ Kd7 4.Qd3+ Kc7 5.Qc4+ Kd7 6.Qb5+ Kc7 7.Qc5+ Kd7 8.Qa7+
  +-  (7.63)   Depth: 17   00:00:03  208kN
1...Kc7 2.Qa5+ Kc6 3.Qc3+ Kd7 4.Qd3+ Kc7 5.Qc4+ Kd7 6.Qb5+ Kc7 7.Qc5+ Kd7 8.Qa7+
  +-  (7.63)   Depth: 18   00:00:04  292kN
1...Kc7 2.Qa5+ Kc6 3.Qc3+ Kd7 4.Qd3+ Kc7 5.Qc4+ Kd7 6.Qb5+ Kc7 7.Qc5+ Kd7 8.Qa7+
  +-  (7.63)   Depth: 19   00:00:06  424kN
1...Kc7 2.Qa5+ Kc6 3.Qc3+ Kd7 4.Qd3+ Kc7 5.Qc4+ Kd7 6.Qb5+ Kc7 7.Qc5+ Kd7 8.Qa7+
  +-  (7.63)   Depth: 20   00:00:08  620kN
1...Kc7 2.Qa5+ Kc6 3.Qc3+ Kd7 4.Qd3+ Kc7 5.Qc4+ Kd7 6.Qb5+ Kc7 7.Qc5+ Kd7 8.Qa7+
  +-  (7.63)   Depth: 21   00:00:12  906kN
1...Kc7 2.Qa5+ Kc6 3.Qc3+ Kd7 4.Qd3+ Kc7 5.Qc4+ Kd7 6.Qb5+ Kc7 7.Qc5+ Kd7 8.Qa7+
  +-  (7.63)   Depth: 22   00:00:18  1351kN
1...Kc7 2.Qa5+ Kc6 3.Qc3+ Kd7 4.Qd3+ Kc7 5.Qc4+ Kd7 6.Qb5+ Kc7 7.Qc5+ Kd7 8.Qa7+
  +-  (7.63)   Depth: 23   00:00:36  2484kN
1...Kc7 2.Qa5+ Kc6 3.Qc3+ Kd7 4.Qd3+ Kc7 5.Qc4+ Kd7 6.Qb5+ Kc7 7.Qc5+ Kd7 8.Qa7+
  +-  (7.63)   Depth: 24   00:00:48  3451kN
1...Kc7 2.Qa5+ Kc6 3.Qc3+ Kd7 4.Qd3+ Kc7 5.Qc4+ Kd7 6.Qb5+ Kc7 7.Qc5+ Kd7 8.Qa7+
  +-  (7.63)   Depth: 25   00:01:08  4963kN

(, AMD 21.05.2007)

Analysis by Rybka 2.2n2 mp 32-bit :

1...Kc7
  +-  (6.82)   Depth: 2   00:00:00
1...Kc7
  +-  (6.82)   Depth: 3   00:00:00
1...Kc7
  +-  (9.21)   Depth: 4   00:00:00
1...Kc7 2.Qa5+
  +-  (9.20)   Depth: 5   00:00:00
1...Kc7 2.Qa5+ Kc8 3.Qc3+ Rc7
  +-  (9.20)   Depth: 6   00:00:00  10kN
1...Kc7 2.Qa5+ Kc8 3.Qc3+ Rc7 4.Qh3+ Kb8
  +-  (9.20)   Depth: 7   00:00:00  12kN
1...Kc7 2.Qa5+ Kc8 3.Qc3+ Rc7 4.Qh3+ Kb8 5.Qg3
  +-  (9.20)   Depth: 8   00:00:00  15kN
1...Kc7 2.Qa5+ Kc8 3.Qc3+ Rc7 4.Qh3+ Kb8 5.Qg3 e1Q+ 6.Qxe1 Rxc2
  +-  (9.20)   Depth: 9   00:00:00  23kN
1...Kc7 2.Qa5+
  +-  (9.20)   Depth: 10   00:00:00  28kN
1...Kc7 2.Qa5+
  +-  (9.20)   Depth: 11   00:00:01  37kN
1...Kc7 2.Qa5+ Kc8 3.Qc3+ Rc7 4.Qh3+ Kb8 5.Qg3 e1Q+ 6.Qxe1 Rxc2 7.Qb4+ Ka7 8.Qa5+
  +-  (9.20)   Depth: 12   00:00:01  55kN
1...Kc7 2.Qa5+ Kc8 3.Qc3+ Rc7 4.Qh3+ Kb8 5.Qg3 e1Q+ 6.Qxe1 Rxc2 7.Qb4+
  +-  (9.20)   Depth: 13   00:00:01  76kN
1...Kc7 2.Qa5+ Kc8 3.Qc3+ Rc7 4.Qh3+ Kb8 5.Qg3 e1Q+ 6.Qxe1 Rxc2 7.Qb4+
  +-  (9.20)   Depth: 14   00:00:01  111kN
1...Kc7 2.Qa5+ Kc8 3.Qc3+ Rc7 4.Qh3+ Kb8 5.Qg3 e1Q+ 6.Qxe1 Rxc2 7.Qb4+
  +-  (9.20)   Depth: 15   00:00:02  155kN
1...Kc7 2.Qa5+ Kc8 3.Qc3+ Rc7 4.Qh3+ Kb8 5.Qg3 e1Q+ 6.Qxe1 Rxc2 7.Qb4+
  +-  (9.20)   Depth: 16   00:00:02  214kN
1...Kc7 2.Qa5+ Kc8 3.Qc3+ Rc7 4.Qh3+ Kb8 5.Qg3 e1Q+ 6.Qxe1 Rxc2 7.Qb4+ Ka7 8.Qa5+
  +-  (9.20)   Depth: 17   00:00:02  307kN
1...Kc7 2.Qa5+ Kc8 3.Qc3+ Rc7 4.Qh3+ Kb8 5.Qg3 e1Q+ 6.Qxe1 Rxc2 7.Qb4+
  +-  (9.20)   Depth: 18   00:00:03  435kN
1...Kc7 2.Qa5+ Kc8 3.Qc3+ Rc7 4.Qh3+ Kb8 5.Qg3 e1Q+ 6.Qxe1 Rxc2 7.Qb4+ Ka7 8.Qa5+
  +-  (9.20)   Depth: 19   00:00:04  623kN
1...Kc7 2.Qa5+ Kc8 3.Qc3+ Rc7 4.Qh3+ Kb8 5.Qg3 e1Q+ 6.Qxe1 Rxc2 7.Qb4+
  +-  (9.20)   Depth: 20   00:00:06  903kN
1...Kc7 2.Qa5+ Kc8 3.Qc3+ Kd7 4.Qd3+ Kc8 5.Qa6+ Kc7 6.Qa7+ Kd6 7.Qd4+ Ke6 8.Qb4
  +-  (9.20)   Depth: 21   00:01:12  10355kN
1...Kc7 2.Qa5+ Kc8 3.Qc3+ Kd7 4.Qd3+ Kc8 5.Qa6+ Kc7 6.Qa7+ Kd6 7.Qd4+ Ke6 8.Qb4
  +-  (9.20)   Depth: 22   00:01:15  10707kN
1...Kc7 2.Qa5+ Kc8 3.Qc3+ Kd7 4.Qd3+ Kc8 5.Qa6+ Kc7 6.Qa7+
  +-  (9.20)   Depth: 23   00:01:21  11484kN
1...Kc7 2.Qa5+ Kc8 3.Qc3+ Kd7 4.Qd3+ Kc8 5.Qa6+ Kc7 6.Qa7+
  +-  (9.20)   Depth: 24   00:01:29  12705kN
1...Kc7 2.Qa5+ Kc8 3.Qc3+ Kd7 4.Qd3+ Kc8 5.Qa6+ Kc7 6.Qa7+
  +-  (9.20)   Depth: 25   00:01:38  13991kN
1...Kc7 2.Qa5+ Kc8 3.Qc3+ Kd7 4.Qd3+ Kc8 5.Qa6+ Kc7 6.Qa7+
  +-  (9.20)   Depth: 26   00:01:53  16130kN
1...Kc7 2.Qa5+ Kc8 3.Qc3+ Kd7 4.Qd3+ Kc8 5.Qa6+ Kc7 6.Qa7+ Kd6 7.Qd4+ Ke6 8.Qb4
  +-  (9.20)   Depth: 27   00:02:13  19075kN
1...Kc7 2.Qa5+ Kc8 3.Qc3+ Kd7 4.Qd3+ Kc8 5.Qa6+ Kc7 6.Qa7+ Kd6 7.Qd4+ Ke6 8.Qb4
  +-  (9.20)   Depth: 28   00:02:42  23518kN
1...Kc7 2.Qa5+ Kc8 3.Qc3+ Kd7 4.Qd3+ Kc8 5.Qa6+ Kc7 6.Qa7+
  +-  (9.20)   Depth: 29   00:03:16  28816kN
1...Kc7 2.Qa5+ Kc8 3.Qc3+
  +-  (9.20)   Depth: 30   00:04:08  36593kN
1...Kc7 2.Qa5+ Kc8 3.Qc3+
  +-  (9.20)   Depth: 31   00:06:10  54874kN
1...Kc7 2.Qa5+ Kc8 3.Qc3+
  +-  (9.20)   Depth: 32   00:08:59  79584kN

(, AMD 21.05.2007)

Analysis by Rybka 2.3.1 32-bit :

1...Kc7
  +-  (7.55)   Depth: 2   00:00:00
1...Kc8
  +-  (7.30)   Depth: 2   00:00:00
1...Kc8
  +-  (7.21)   Depth: 3   00:00:00
1...Kc8
  +-  (7.10)   Depth: 4   00:00:00
1...Kc8 2.Qc6+ Kd8
  +-  (6.99)   Depth: 5   00:00:00
1...Kc8 2.Qc6+ Kd8 3.Qd6+ Ke8
  +-  (7.30)   Depth: 6   00:00:00  3kN
1...Ke8 2.Qb5+ Kf7 3.Qc4+ Re6
  =  (0.00)   Depth: 6   00:00:00  6kN
1...Ke8 2.Qb5+ Kf7 3.Qc4+ Re6 4.Qf4+ Rf6
  =  (0.00)   Depth: 7   00:00:00  6kN
1...Ke8 2.Qb5+ Kf7 3.Qc4+ Re6 4.Qf4+ Rf6 5.Qc7+ Kg6
  =  (0.00)   Depth: 8   00:00:00  7kN
1...Ke8 2.Qb5+ Kf7 3.Qc4+ Re6 4.Qf4+ Rf6 5.Qc7+ Kg6 6.Qg3+ Kf7
  =  (0.00)   Depth: 9   00:00:00  7kN
1...Ke8 2.Qb5+ Kf7 3.Qc4+ Re6 4.Qf4+ Rf6 5.Qc7+ Kg6 6.Qg3+ Kf7 7.Qc7+ Kg6 8.Qg3+
  =  (0.00)   Depth: 10   00:00:01  8kN
1...Ke8 2.Qb5+ Kf7 3.Qc4+ Re6 4.Qf4+ Rf6 5.Qc7+ Kg6 6.Qg3+ Kf7 7.Qc7+ Kg6 8.Qg3+
  =  (0.00)   Depth: 11   00:00:01  10kN
1...Ke8 2.Qb5+ Kf7 3.Qc4+ Re6 4.Qf4+ Rf6 5.Qc7+ Kg6 6.Qg3+ Kf7 7.Qc7+ Kg6 8.Qg3+
  =  (0.00)   Depth: 12   00:00:01  12kN
1...Ke8 2.Qb5+ Kf7 3.Qc4+ Re6 4.Qf4+ Rf6 5.Qc7+ Kg6 6.Qg3+ Kf7 7.Qc7+ Kg6 8.Qg3+
  =  (0.00)   Depth: 13   00:00:02  15kN
1...Ke8 2.Qb5+ Kf7 3.Qc4+ Re6 4.Qf4+ Rf6 5.Qc7+ Kg6 6.Qg3+ Kf7 7.Qc7+ Kg6 8.Qg3+
  =  (0.00)   Depth: 14   00:00:02  20kN
1...Ke8 2.Qb5+ Kf7 3.Qc4+ Re6 4.Qf4+ Rf6 5.Qc7+ Kg6 6.Qg3+ Kf7 7.Qc7+ Kg6 8.Qg3+
  =  (0.00)   Depth: 15   00:00:02  29kN
1...Ke8 2.Qb5+ Kf7 3.Qc4+ Re6 4.Qf4+ Rf6 5.Qc7+ Kg6 6.Qg3+ Kf7 7.Qc7+ Kg6 8.Qg3+
  =  (0.00)   Depth: 16   00:00:02  43kN
1...Ke8 2.Qb5+ Kf7 3.Qc4+ Re6 4.Qf4+ Rf6 5.Qc7+ Kg6 6.Qg3+ Kf7 7.Qc7+ Kg6 8.Qg3+
  =  (0.00)   Depth: 17   00:00:03  63kN
1...Ke8 2.Qb5+ Kf7 3.Qc4+ Re6 4.Qf4+ Rf6 5.Qc7+ Kg6 6.Qg3+ Kf7 7.Qc7+ Kg6 8.Qg3+
  =  (0.00)   Depth: 18   00:00:03  94kN
1...Ke8 2.Qb5+ Kf7 3.Qc4+ Re6 4.Qf4+ Rf6 5.Qc7+ Kg6 6.Qg3+ Kf7 7.Qc7+ Kg6 8.Qg3+
  =  (0.00)   Depth: 19   00:00:03  141kN
1...Ke8 2.Qb5+ Kf7 3.Qc4+ Re6 4.Qf4+ Rf6 5.Qc7+ Kg6 6.Qg3+ Kf7 7.Qc7+ Kg6 8.Qg3+
  =  (0.00)   Depth: 20   00:00:04  213kN
1...Ke8 2.Qb5+ Kf7 3.Qc4+ Re6 4.Qf4+ Rf6 5.Qc7+ Kg6 6.Qg3+ Kf7 7.Qc7+ Kg6 8.Qg3+
  =  (0.00)   Depth: 21   00:00:04  322kN
1...Ke8 2.Qb5+ Kf7 3.Qc4+ Re6 4.Qf4+ Rf6 5.Qc7+ Kg6 6.Qg3+ Kf7 7.Qc7+ Kg6 8.Qg3+
  =  (0.00)   Depth: 22   00:00:06  489kN
1...Ke8 2.Qb5+ Kf7 3.Qc4+ Re6 4.Qf4+ Rf6 5.Qc7+ Kg6 6.Qg3+ Kf7 7.Qc7+ Kg6 8.Qg3+
  =  (0.00)   Depth: 23   00:00:09  740kN
1...Ke8 2.Qb5+ Kf7 3.Qc4+ Re6 4.Qf4+ Rf6 5.Qc7+ Kg6 6.Qg3+ Kf7 7.Qc7+ Kg6 8.Qg3+
  =  (0.00)   Depth: 24   00:00:14  1119kN
1...Ke8 2.Qb5+ Kf7 3.Qc4+ Re6 4.Qf4+ Rf6 5.Qc7+ Kg6 6.Qg3+ Kf7 7.Qc7+ Kg6 8.Qg3+
  =  (0.00)   Depth: 25   00:00:20  1654kN
1...Ke8 2.Qb5+ Kf7 3.Qc4+ Re6 4.Qf4+ Rf6 5.Qc7+ Kg6 6.Qg3+ Kf7 7.Qc7+ Kg6 8.Qg3+
  =  (0.00)   Depth: 26   00:00:30  2519kN
1...Ke8 2.Qb5+ Kf7 3.Qc4+ Re6 4.Qf4+ Rf6 5.Qc7+ Kg6 6.Qg3+ Kf7 7.Qc7+ Kg6 8.Qg3+
  =  (0.00)   Depth: 27   00:00:49  4119kN
1...Ke8 2.Qb5+ Kf7 3.Qc4+ Re6 4.Qf4+ Rf6 5.Qc7+ Kg6 6.Qg3+ Kf7 7.Qc7+ Kg6 8.Qg3+
  =  (0.00)   Depth: 28   00:01:35  7594kN

(, AMD 21.05.2007)

Regards,
Gambitto.
Parent - - By Vasik Rajlich (Silver) [hu] Date 2007-05-21 10:23
This is true, but the bugs in 2.2n2 were less serious.

In fact, the day before the tournament, two reliable testers (Majd and Bjorn) warned me that there was something fishy in the multi-processor version, while another tester sent me a very poor blitz result for the MP version (compared to my SP results). We have some lessons-learned from this tournament, no question about that.

Vas
Parent - By Kapaun (****) [de] Date 2007-05-21 11:51
Okay, so far for 2.3.2 coming out soon, sniff....
Parent - - By Shaun (****) [gb] Date 2007-05-21 13:05
A preview of the CCRL 40/4 results that will be uploaded tonight

  Rybka 2.3.1 64-bit 2CPU                 =>  3078 ( +24  -23)  80.5% -233.5 (233.5)  23.4%   798 (+549-62=187)
  Rybka 2.1d3 64-bit 2CPU                 =>  3064 ( +27  -26)  79.7% -223.3 (223.3)  24.5%   607 (+409-49=149)
  Rybka 2.2 64-bit 2CPU                   =>  3058 ( +13  -13)  74.2% -180.6 (180.6)  28.8%  2513 (+1502-286=725)
  Rybka 2.3.2 Beta 64-bit 2CPU            =>  3054 ( +20  -20)  75.0% -179.2 (179.5)  32.4%   977 (+574-86=317)
  Rybka 2.2n2 64-bit 2CPU                 =>  3048 ( +24  -23)  77.0% -198.3 (198.3)  26.9%   763 (+485-73=205)
  Rybka 2.3 LK 64-bit 2CPU                =>  3047 ( +24  -24)  77.7% -208.2 (208.2)  25.0%   736 (+480-72=184)
  Rybka 2.1o 64-bit 2CPU                  =>  3046 ( +24  -23)  76.5% -195.3 (196.0)  28.6%   735 (+457-68=210)


even broken 2.3.2 is not that bad :lol:

The margin of error would allow 2.3.2 to be the strongest version just doubtful - looking forward to a fixed version.

For info the 977 Rybka 2.3.2 beta games were run on 2 Core 2 Duo machines over 3 days finishing this morning.

Shaun
Parent - - By Dragon Mist (****) [hr] Date 2007-05-21 13:09
Oh no, not another non-best Rybka!!

:-(

Dragon MIst
Parent - By Shaun (****) [gb] Date 2007-05-21 13:29
To be expected when a bug was found - the danger of testing betas

Shaun
Parent - - By Uri Blass (*****) [il] Date 2007-05-21 13:23
I think that it may be interesting to have rybka2.3.2 beta 64 bits 1 cpu in the list.
Maybe it can get first place when 2.3.1 2 cpu will only get second place.

Uri
Parent - - By Shaun (****) [gb] Date 2007-05-21 13:42
I think I will wait :lol:

P.S. when are we going to see an mp version of Movei???
Parent - By Uri Blass (*****) [il] Date 2007-05-21 13:51
MP is not in my plans for the near future.

Uri
Parent - By Vasik Rajlich (Silver) [hu] Date 2007-05-23 07:12
Probably so, but not for long :)

Vas
Parent - By noctiferus (***) [it] Date 2007-05-20 20:17 Edited 2007-05-20 20:22
First, Vas, congratulations.

Then, to all: now at

http://www.csvn.nl/index.php?option=com_content&task=view&id=258&Itemid=28

the pgn is complete, even if the tag says "download rounds 1-7"
Parent - - By Gaмßito (****) [cr] Date 2007-05-20 20:28
Vas,

Congratulations for this nice victory!

At the end of this Tournament, Rybka has closed with fastener of gold with three consecutive victories!

I enjoyed the last game. It seemed that Karpov was playing it.

Regards,
Gambitto.
Parent - By Felix Kling (Gold) [de] Date 2007-05-20 20:51
I also like this game. It's a pefect positional game, white has a very good knight, the opponent a weak bishop, black has two weak pawns on d6 and also on b6. with rook c3-c6 white attacks the second weakness and black loses. Very nice. I will show this game to my pupils!
Parent - - By hafi (*) [de] Date 2007-05-20 21:11
If there is a bug, as identified in the game of round 6, will it be corrected in the soon to be released update.

Anyhow,

Congratulations

P.S: Little has been about the items being fixed in the update. Will it resolve some of the Zugzwang issues. (I am aware that these do not contribute much to the overall Elo rating, but still they are a nuisance.
Parent - - By Gaмßito (****) [cr] Date 2007-05-21 02:45
Yes,  problems with Zugzwang positions are very important to be fixed as soon as possible. These positions are a bit strange, but these are presented from time to time and indeed are quite important.

I suppose Vas will not fix this problem in the next 2.3.2 version, but surely for the 3.0 version.

Regards,
Gambitto.
Parent - By grolich (***) Date 2007-05-21 09:55
Solving the zugzwang problem revolves around one issue: When to test for the validity of a null move fail high.
Vas has claimed in the past that it IS possible to always test for the validity, but it is an ugly solution (with a clear downside in calculation performance). So it will eventually boil down to: will Vas find a way to dramatically improve the heuristic which controls the decision of when to run the null move fail high validity test (the null move test is, funnily enough, a heuristic itself. So we have a heuristic to decide when to run a test which validates the correctness of another heuristic. Sounds complex enough? If not, you probably do not understand it sufficiently) or will he decide to temporarily use the "brute force" method.

I don't believe the latter option will be used, unless Vas encounters a serious coder's (or algorithmic) block... So we'll have to wait and see.
Coming up with those heuristics is generally a lengthy research into what actually works. Which means it will probably only be dealt with after more important issues.
Parent - - By noctiferus (***) [it] Date 2007-05-20 21:19
Just a marginal question about the rules: was it possible to retire a move due to a misclick by the operator?
Not important, but it could be an explanation to something I saw looking at a game.
Parent - By Carl Bicknell (*****) [gb] Date 2007-05-20 21:21
EXCELLENT result Vas & team - just what I was hoping for!

WELL DONE
Parent - - By Harvey Williamson (*****) Date 2007-05-20 21:24
Yes it is possible as what counts is the position on the real board and time on the real Chess clock.

The atmosphere is very relaxed and a mouse slip would not be a problem.

hmm just noticed my flag has changed to Dutch?! Must be the Leiden effect.
Parent - By noctiferus (***) [it] Date 2007-05-20 21:26
OK. I was just trying to explain to myself a sudden drop and  sudden increase of the eval. Thanks so much
Parent - - By Vasik Rajlich (Silver) [hu] Date 2007-05-21 10:20
Thanks for all of the comments.

We used Rybka 2.3.2 Beta in rounds 1-6 and Rybka 2.2n2 in rounds 7-9. There are a number of things which need to be fixed in 2.3.2 Beta before it's ready for release. The hash bug from the game vs Sjeng is obviously one of them. It shouldn't take a massive amount of time, though.

Re. Zugzwang, Rybka 2.3.2 is a little bit better there, I did piddle with it a little but, it's not a completely solution.

Vas
Parent - - By Dragon Mist (****) [hr] Date 2007-05-21 10:42
Wow! Now this tells a lot about 2.3, 2.3 LK and 2.3.1 versions!!

Dragon Mist
Parent - - By Michael Waesch [de] Date 2007-05-21 11:44
LOL. Now may be quite a few will revert to 2.2n2 ...

Mike
Parent - - By FWCC (***) [us] Date 2007-05-21 12:42
Congratulations on a fine victory.Rybka is still number 1,good work Rybka team!!
Parent - By FICGS (**) [fr] Date 2007-05-22 02:20
Congrats. Nice result ! .. No doubt that Rybka is the best chess engine... Waiting for Fritz 12 :)
Parent - By Hetman (*****) Date 2007-05-24 16:38
Hello,
congratulations to the Rybka team. 
Rybka advantage over the field is similair to that which did have R.J. Fischer in 1970's over the concurrence.
Rybka is a Fish , Fischer is Rybak :-).

Regards
Hetman
Up Topic Rybka Support & Discussion / Rybka Discussion / Rybka @ Leiden ICT7

Powered by mwForum 2.27.4 © 1999-2012 Markus Wichitill