Not logged inRybka Chess Community Forum
Up Topic Rybka Support & Discussion / Rybka Discussion / 29th ODCCC 2009 in Leiden
1 2 3 Previous Next  
- - By Lukas Cimiotti (Bronze) [de] Date 2009-10-16 15:45
First Rybka game:

[Event "29th ODCCC'"]
[Site "?"]
[Date "2009.10.16"]
[Round "?"]
[White "Hermann"]
[Black "Rybka 3t88 cluster"]
[Result "0-1"]
[ECO "E10"]
[PlyCount "92"]

1. Nf3 Nf6 2. c4 e6 3. d4 Nc6 4. a3 d6 5. Nc3 g6 6. e4 Bg7 7. h3 O-O 8. Bg5 h6
9. Be3 Re8 10. Bd3 Nd7 11. O-O e5 12. d5 Nd4 13. Bb1 c5 14. Bd3 b6 15. Nd2 Nf6
16. f4 exf4 17. Bxf4 a6 18. Rf2 Qe7 19. a4 Nd7 20. a5 Ne5 21. Bf1 bxa5 22. Rxa5
Rb8 23. Ra2 h5 24. Qb1 g5 25. Be3 g4 26. Bxd4 cxd4 27. Nd1 g3 28. Re2 Bg4 29.
Re1 Bh6 30. Qc2 Bxh3 31. Be2 d3 32. Bxd3 Bxd2 33. Qxd2 Qh4 34. Ne3 Bg4 35. Nf1
Nf3+ 36. gxf3 Bxf3 37. Qg2 Bxg2 38. Kxg2 Rb3 39. Ra3 Rxb2+ 40. Re2 Rxe2+ 41.
Bxe2 Qxe4+ 42. Bf3 Qc2+ 43. Kxg3 h4+ 44. Kxh4 Qxc4+ 45. Kg3 Qxf1 46. Bg2 Qg1
0-1
Parent - - By Bouddha (****) [ch] Date 2009-10-16 16:05
Excellent game thank you !

That was an interesting way to play the opening.
Out of expected theory in a position which is not worse with plenty of possibilities to play for a win !
Good strategy.

Thanks for posting all games of the cluster.

regards
Parent - By Nelson Hernandez (Gold) [us] Date 2009-10-17 14:06
Actually Hermann played that opening surprisingly well, I thought.  It got caught in the middle-game.  I'm showing 13...c5 as the novelty move.
Parent - - By Highendman (****) Date 2009-10-16 16:36
What's the schedule? Can we see them live anywhere?
Congrats on a fine win!
Parent - - By Eelco de Groot (***) Date 2009-10-16 18:01
There are no live games, I'm afraid. Last year I believe Harvey could be followed on Playches but he is not playing now. The remaining schedule for today is shortened a bit because of the delays, same number of games but only 75 minutes per game. Rybka is free in the second round.

29th ODCCC Round 2       
Friday, 16 October 2009 16:39 
Hansdamf - Hermann

Deep Sjeng - The King

Pandix - Joker

Ktulu - The Baron

Free round : Rybka


Planned time schedule is here barring any more router problems.

The games are made available after the rounds though and there are more pictures already so we can take in the atmosphere of the tournament!

Parent - By Highendman (****) Date 2009-10-16 18:05
Many thanks for the update and pictures!
Parent - - By Permanent Brain (*****) Date 2009-10-16 19:38
Round 3 has Deep Sjeng - Rybka:

[Event "29th Open Dutch Computer Chess Champion"]
[Site "?"]
[Date "2009.10.16"]
[Round "3"]
[White "Deep Sjeng X"]
[Black "Rybka X"]
[Result "0-1"]
[Annotator "€rdo"]
[PlyCount "89"]

1. e4 c5 2. Nf3 e6 3. d4 cxd4 4. Nxd4 a6 5. Nc3 Qc7 6. Be2 b5 7. O-O Bb7 8. Re1
Nc6 9. Nxc6 dxc6 10. e5 Rd8 11. Bd3 c5 12. Qe2 c4 13. Be4 Bxe4 14. Qxe4 Ne7 15.
Be3 b4 16. Nd1 Nd5 17. Bg5 Rd7 18. Ne3 h6 19. Nxd5 Rxd5 20. Bh4 c3 21. Rad1 g5
22. Rxd5 exd5 23. Qxd5 gxh4 24. e6 Be7 25. exf7+ Kf8 26. bxc3 bxc3 27. h3 Kg7
28. Qd4+ Bf6 29. Qg4+ Kxf7 30. Qe6+ Kg7 31. Qg4+ Kf8 32. Qf3 Qf7 33. Re4 h5 34.
Rb4 Be5 35. Qd3 Kg7 36. Rxh4 Qxa2 37. g3 Qb1+ 38. Kg2 Qb7+ 39. Kg1 Qc6 40. Rc4
Qf6 41. Qd7+ Qf7 42. Qxf7+ Kxf7 43. f4 Bf6 44. Kf2 Rd8 45. Rc7+ 0-1

3r4/2R2k2/p4b2/7p/5P2/2p3PP/2P2K2/8 b - - 0 45
Parent - - By Arifur (**) [gb] Date 2009-10-16 19:43
This was 72 cores vs 72 cores, correct?
Parent - - By Lukas Cimiotti (Bronze) [de] Date 2009-10-16 20:15
Well, we had / have 72 cores. I don't know about Sjeng. I asked Erdo. He told me they could use 288 + 48 cores at max. But he didn't tell me how many cores they really used. I hope GCP will tell us.
Parent - - By suj (***) Date 2009-10-16 20:53
Sjeng ran on 64 cores in total
Parent - - By Lukas Cimiotti (Bronze) [de] Date 2009-10-16 20:55
Thanks for that information :)
Parent - By GCP (***) [nl] Date 2009-10-17 07:45
It might vary a bit throughout the entire tournament. I did updates to clustering shortly before the tournament and we're still looking for the optimal settings+hw setup. Unfortunately we didn't meet Rybka in the last round :) But I'm happy with what I saw so far.

NPS varies from 45 to 85Mnps.
Parent - By GCP (***) [nl] Date 2009-10-17 07:49

>I hope GCP will tell us.


You have to check with suj, I'm only watching the master (4x6 2.4Ghz AMD) and he sets up the rest.
Parent - By Lukas Cimiotti (Bronze) [de] Date 2009-10-16 20:13
Thanks for posting that game - saves me some work :)
Parent - By Felix Kling (Gold) [de] Date 2009-10-16 23:43
And a picture:
Parent - By buffos (Silver) [gr] Date 2009-10-17 08:45
I would be very interesting to know the evaluations of both engines after the sacr. Its more important than the actual result. If both engines thought it was nice, its another story.
Its also not clear from first sight if white is losing or he has a draw.
Parent - - By Felix Kling (Gold) [de] Date 2009-10-16 22:44
Pictures are really great :) Last time I unsuccessfully tried to get one for our website, this time it should be easy :) Who took the photos? Before I can publish them on our website I would like to make sure this is ok.
Parent - - By Eelco de Groot (***) Date 2009-10-16 23:18
Hi Felix,

Yes, it's almost as if you are there, participating :) I sure would have liked some of those frites and hamburger that Johan de Koning was having for a meal. Regular healthfood it is not but for weather like we're having today with all the rain and wind, nothing better.

I think the pictures are probably made by the tournament director Jan Krabbenbos, I'm pasting the link of the contactform in here, he's your webmaster-colleague on the computerschaak site.

Regards, Eelco
Parent - By Permanent Brain (*****) Date 2009-10-17 00:17
If someone from the CSVN website reads this:

The entry for round 1 has disappeared.
Parent - - By Eelco de Groot (***) Date 2009-10-18 09:15 Edited 2009-10-18 09:18
Kibitzing, I think the Baron, on Harvey's machine, has chances to reach a draw at the moment:

[Event "?"]
[Site "?"]
[Date "????.??.??"]
[Round "?"]
[White "Rybka"]
[Black "The Baron"]
[Result "*"]
[Annotator "Nightwish35"]
[PlyCount "56"]

1. d4 d5 2. c4 dxc4 3. e3 e5 4. Bxc4 exd4 5. exd4 Bd6 6. Nf3 Nf6 7. O-O O-O 8.
h3 Nc6 9. Nc3 h6 10. Qc2 Na5 11. Bd3 Re8 12. Bd2 Nc6 13. a3 Be6 14. Rfe1 Qd7
15. b4 Rad8 16. Ne4 Nxe4 17. Bxe4 f5 18. Bd3 Kh8 19. d5 Bxd5 20. Bxf5 Qf7 21.
Bg6 Rxe1+ 22. Rxe1 Qf8 23. Be4 Qf7 24. Bc3 Re8 25. b5 Nd8 26. Nh4 Bxe4 27. Rxe4
Ne6 28. Rg4 h5 *

Rybka to move:

4r2k/ppp2qp1/3bn3/1P5p/6RN/P1B4P/2Q2PP1/6K1 w - -


Engine: Ancalagon 1.3 Weak Squares 180 Board Control middlegame 50 endgame 50
Build 357 (Athlon 2009 MHz, 256 MB) by Romstad, Costalba, Kiiski, de Groot

2.00   0:06   -0.12    29.Ng6+ Kg8 30.Ra4 Ng5 31.Bd2 Qd5
                       32.Nf4 Bxf4 33.Rxf4 Re2 (3.171.850) 475

2.03   0:12   -0.04    29.Rc4 Nf4 30.Re4 Rxe4 31.Qxe4 Qe6
                       32.Qxe6 Nxe6 33.a4 Kh7 34.Nf5 (5.766.824) 465

2.06   0:14   +0.02    29.Re4 Nf4 30.Nf5 Rxe4 31.Qxe4 Ne2+
                       32.Qxe2 Qxf5 33.a4 Qb1+ 34.Qe1 Qxe1+
                       35.Bxe1 Kh7 36.Bc3 (6.954.838) 468

3.01   0:21   +0.02    29.Re4 Nf4 30.a4 Rxe4 31.Qxe4 Qb3
                       32.Be5 Qd1+ 33.Kh2 Bxe5 34.Qxe5 Qxa4
                       35.Qxc7 (10.550.156) 489

4.01   0:30   +0.16    29.Re4 Nf4 30.a4 Rxe4 31.Qxe4 Kg8
                       32.Be5 Qe6 33.Nf3 Bxe5 34.Nxe5 Ng6
                       35.Qxg6 Qxe5 (15.104.744) 492

5.01   0:54   +0.38    29.Re4 Nf4 30.a4 Rxe4 31.Qxe4 Kg8
                       32.Be5 Bxe5 33.Qxe5 Qc4 34.g3 Ne6
                       35.Qxh5 Qxa4 36.Qe8+ Nf8 37.Qc8 Qxb5
                       38.Qxc7 (27.052.328) 499

6.01   1:12   +0.12    29.Re4 Nf4 30.Nf5 Rxe4 31.Qxe4 Ne2+
                       32.Qxe2 Qxf5 33.Qb2 Qg5 34.a4 Kg8
                       35.g3 h4 36.Bd2 Qf5 37.Qb3+ Qf7
                       38.Qe3 hxg3 39.fxg3 (36.410.314) 504

7.01   1:46   +0.33    29.Re4 Nf4 30.a4 Kg8 31.Rxe8+ Qxe8
                       32.Nf5 Ne2+ 33.Kf1 Nxc3 34.Qxc3 Qe5
                       35.Qc4+ Kf8 36.Nxd6 Qxd6 37.Ke2 Qe5+
                       38.Kd3 Ke7 39.Qe4 Qxe4+ 40.Kxe4 g6 (53.492.704) 502

8.01   5:22   +0.39    29.Re4 Nf4 30.a4 Kg8 31.Rxe8+ Qxe8
                       32.Nf5 Ne2+ 33.Kf1 Nxc3 34.Qxc3 Qe5
                       35.Qb3+ Kh7 36.Nxd6 Qxd6 37.Qf3 g6
                       38.Ke2 Qe6+ 39.Kd1 Qb6 40.Kc2 (151.450.520) 469
Parent - - By Eelco de Groot (***) Date 2009-10-18 09:27
Ancalagon predicts a draw, Rybka to play move 33:

4q1k1/ppp3p1/8/1P5p/1P3n1N/7P/2Q2PP1/6K1 w - -


Engine: Ancalagon 1.3 Weak Squares 180 Board Control middlegame 50 endgame 50
Build 357 (Athlon 2009 MHz, 256 MB) by Romstad, Costalba, Kiiski, de Groot

2.00   0:02   +0.25    33.Qa2+ Kh7 34.Qc2+ Kg8 35.Nf3 Qxb5
                       36.Qxc7 Qxb4 37.Qb8+ Kf7 38.Qxa7 g6
                       39.Qe3 Qb1+ 40.Kh2 (1.828.552) 692

3.01   0:04   +0.18    33.Qa2+ Kh7 34.Qc2+ Kg8 35.Nf3 Qxb5
                       36.Qxc7 Qxb4 37.Qb8+ Kf7 38.Qxa7 Ne2+
                       39.Kf1 Nc3 40.Qe3 Qb1+ 41.Qe1 Qb5+
                       42.Kg1 (2.945.873) 698

4.01   0:06   +0.06    33.Qa2+ Kh7 34.Qc2+ Kg8 35.Nf3 Qxb5
                       36.Qxc7 Qxb4 37.Qb8+ Kf7 38.Qxa7 g6
                       39.Qa1 b5 40.Ng5+ Kg8 41.Qe5 Qb1+
                       42.Kh2 (4.759.264) 698

5.01   0:17   +0.26    33.Qa2+ Kh7 34.Qc2+ Kg8 35.Qxc7 Qe1+
                       36.Kh2 Qxf2 37.Qe7 Kh7 38.Qg5 Qd4
                       39.Ng6 Nxg6 40.Qxh5+ Kg8 41.Qxg6 Qe5+
                       42.Qg3 Qxb5 43.Qb8+ Kf7 44.Qxa7 Qxb4
                       45.Kg3 (12.287.454) 700

6.01   0:37   +0.16    33.Qa2+ Kh7 34.Qc2+ Kg8 35.Qc4+ Qf7
                       36.Qe4 Nd5 37.Qd4 Nb6 38.Nf3 Qd7
                       39.Qa1 Nc8 40.Qa2+ Qf7 41.Qa4 Qf4
                       42.Qc2 (25.942.335) 695

7.01   1:12   +0.07    33.Qa2+ Kh7 34.Qc2+ Kg8 35.Nf3 Qxb5
                       36.Qxc7 Qxb4 37.Qb8+ Kf7 38.Qxa7 Ne2+
                       39.Kf1 Nf4 40.Qa2+ Ke7 41.Qc2 Kf6
                       42.Qd1 Ke6 43.Nd4+ Ke5 44.Ne2 (49.454.781) 686

7.20   1:58   +0.22    33.Qc4+ Qf7 34.Qe4 Nd5 35.Nf3 g6
                       36.Nh4 Nf4 37.Kh2 b6 38.Nf3 Kg7
                       39.Ne5 (78.574.833) 663

8.01   2:04   +0.21    33.Qc4+ Qf7 34.Qe4 Nd5 35.Qd4 Nb6
                       36.Nf3 Qe7 37.Qd2 Kf8 38.Qa2 Qxb4
                       39.Qxa7 (82.739.549) 664

9.01   3:59   +0.15    33.Qc4+ Qf7 34.Qe4 Nd5 35.Qd4 Nb6
                       36.Nf3 Qc4 37.Qd8+ Kh7 38.Qg5 g6
                       39.Ne5 Qe6 (151.836.510) 634
Parent - By Eelco de Groot (***) Date 2009-10-18 09:46
6k1/pp4p1/2q5/Q2n3p/1P5P/5NP1/5P2/6K1 b - -


Ancalagon not agreeing with all the moves but now 38... a6 should still draw:

10.01  2:40   -0.34    38...a6 39.Qd8+ Kh7 40.Kh2 b6 41.Nd4 Qc4
                       42.Ne6 Qc6 43.Nf8+ Kg8 44.Nd7+ Kh7
                       45.Qe8 Nxb4 46.Qxh5+ Kg8 47.Ne5 (105.594.074) 656
Parent - By oudheusa (*****) [de] Date 2009-10-16 17:22
14. Bd3?! looks strange.

What would the cluster have played after the more logical 14.dc6x ?
Parent - By InspectorGadget (*****) [za] Date 2009-10-16 17:28
Nice game Lukas, my kind of games :)
Parent - - By M ANSARI (*****) [kw] Date 2009-10-16 21:32
An excellent game by DS, and an even more impressive game by Rybka.  Going through that game I had a feeling that black was somehow underdeveloped and that white would crash in due to not being able to castle.  I am not sure where white went wrong, it did feel like white needed to sacrifice some material to take advantage of black's lack of development, but I think it came way too late.  This looks like very sharp opening preparation by DS, and it was very impressive how Rybka fended off any budding initiative and mopped up after DS had over reached.  Very good game from both sides.
Parent - - By Permanent Brain (*****) Date 2009-10-16 23:29 Edited 2009-10-16 23:32
I was too late to watch it live from the start. I wonder what the last book moves were, each... 12.Qe2 stands out. The "normal" and rather old books I have wouldn't play it or don't contain it; more frequent continuations are Qg4 and Qh5.

12.Qh5 is the tournament move in Fritz7.ctg but has a "?" in the Shredder11.ctg. :grin: So, the opinions seem to differ here a lot. Maybe it's engine dependant, or the assessment of a certain continuation has changed.

The positions and moves around the 20th move seem most interesting, 20...c3 21.Tad1 (maybe slightly doubtful?) etc. after which Black remains with a bishop for two pawns, with totally disrupted pawns but as it seems, White doesn't have enough compensation. - I get the impression that it was played quite aggressively by both, e.g. 21...g5 right away instead of inserting 21...Rxd1 22.Rxd1. But it would require deeper analysis if that was really a good alternative. Deep Sjeng did not refute 21...g5 anyway, I just mean to say Rxd1 seems less risky, on a first glance.
Parent - - By GCP (***) [nl] Date 2009-10-17 07:33 Edited 2009-10-17 07:39
Rybka went positive after finding c3!. Deep Sjeng failed low when searching Tad1. I had the impression it was already downhill from there - Rybka pondered almost every move so I'm not sure if there was really a lot white could still do from then on. I'm curious if around Nxd5 we didn't went wrong - had no time to analyse yet.
Parent - - By Highendman (****) Date 2009-10-17 07:53
My book had 15. a3 instead of moving the bishop, to stop the subsequent b4 etc. Did Sjeng consider that move?
Parent - By GCP (***) [nl] Date 2009-10-17 10:38
No, it thinks white has no more than a draw after that. (The line it gives doesn't look very drawish to me, but what do I know... :-P)

For the Rybka team this was theory up to white playing Nd1 (Ne2 expected).
Parent - - By suj (***) Date 2009-10-17 11:36
yes sjeng considered a3 and changed to nd1
Parent - By Highendman (****) Date 2009-10-17 13:47
Thanks. Interesting to see how Rybka would have tried to win after a3. I analysed this opening few months ago, and reached the conclusion that a3 was White's best answer, and should lead to a draw - so I removed the entire line from my 'mental' book for both sides.
Parent - - By Nelson Hernandez (Gold) [us] Date 2009-10-17 14:11
I have the novelty move here as 16...Nd5.  16.Nd1 comes from Ares03, Shredder 8-Pawn_Killer, Deep Fritz 8, 2004.03.23.
Parent - - By Highendman (****) Date 2009-10-17 15:54
May I ask what your stats show for 15. a3?
Parent - - By Nelson Hernandez (Gold) [us] Date 2009-10-17 18:40
My stats show the novelty game as Highendman, Rygka 3 960 - Jarechu, Rybka 3 Dynamic playing to a draw 2009.04.30 and the same line followed only two days later in Jernjoffen, Naum 3.1 - Barsteward, Fritz 11, a white win.  That's all I have.  I have three other choices for white's 15th move, Bg5, Bf4 and a4.
Parent - - By Highendman (****) Date 2009-10-17 18:42
Your database is nothing short of incredible.
Thanks for taking the time to look this up and post.
Parent - - By Nelson Hernandez (Gold) [us] Date 2009-10-17 20:02
The database: years of mindless and repetitive toil.  Somebody has got to do it, otherwise the accumulated knowledge locked up in millions of games will, over time, disappear as hard drives are retired and old files deleted.  I am not really a game archivist, though.  I am more of a position archivist.
Parent - - By Highendman (****) Date 2009-10-17 20:06
Maybe one day you'll open it up for the greater good :)

On a less altruistic note: can you share some stats on your db? i.e. how many quality games you've included in it so far? is it only engine/engine or do you also include human games? I guess you have some decent backup process in place after all the toil that went into it.
Parent - By Nelson Hernandez (Gold) [us] Date 2009-10-17 23:52
Maybe one day I will, but it is not yet on the horizon.

Stats: I am persistently shy on numerical specifics as I don't want to give anyone a number to compete against or that will drive them mad with envy.  However I do collect both human and engine assiduously, and of course I back up constantly, not just on multiple computers but multiple locations (what if there was a fire or I suddenly died?). 

As for the quality games question, I don't know how to answer that.  Define a "quality" game in a way that will endure let's say five years.  What exactly is quality?  My concept of quality is this: how much does the entire length of the game deviate from perfect play?  That's a good way to look at it, I think, except that nobody has access to 32-man TBs so you really don't know with very high precision which games have the highest quality, and even if you could accurately quantify this, what would you include and exclude?  And how would you grade errors in the opening versus errors in the middle or ending?  Rather than try to answer such cosmic questions I accept a fairly large subset of the total that I collect, make improvements to the obvious errors and let numbers and time gradually reveal their own truth (which is never 100% definitive until you see a forced mate or draw).  In the end it's an opening book based on a database.  It's isn't the Oracle at Delphi.
Parent - - By GCP (***) [nl] Date 2009-10-17 10:28
Rd8 put us out of book. Nd1 put Rybka out of book.
Parent - - By Jiri Dufek (**) [cz] Date 2009-10-17 10:46 Edited 2009-10-17 10:48
After all i don't think that this can be counted as book-win. Position is hard (but equal - this doesn't mean drawish) for engines and this time Rybka was happier with black position.
I think that Sjeng last book move was 14.Qxe4 - isn't it true?
Parent - By Felix Kling (Gold) [de] Date 2009-10-17 20:32
btw., if it's no big problem for you, could you or Vas, if you later on comment on the games, make the comments in pgn format so that I can add the games with comments to our cluster games database easier? I can then also export the games with Aquarium's blog export so that people can replay the games here in the forum. I think it would be cool to have a database with commented Rybka games so that those comments aren't "lost" in the forum after some time.
Parent - By GCP (***) [nl] Date 2009-10-18 10:15
We were out of book after Rd8 as I already said.
Parent - By Permanent Brain (*****) Date 2009-10-17 00:43
Great! With 28...Bg4!, Rybka started a fireworks of brilliant moves (30...Bxh3!, 31...d3!, 35...Sf3+!).

1r2r1k1/4qpb1/p2p4/3Pn2p/2PpP1b1/6pP/RP1NR1P1/1Q1N1BK1 w - - 0 29

(after 28...Bg4)
Parent - - By buffos (Silver) [gr] Date 2009-10-17 08:39
Entertaining game. If black was Human, everyone would be A.M.A.Z.E.D

Crushing :)
Parent - - By Maitra (**) [in] Date 2009-10-17 13:31
can you plz post this game?
Parent - By Felix Kling (Gold) [de] Date 2009-10-17 21:35
look above :) it's the first post
Parent - By M ANSARI (*****) [kw] Date 2009-10-18 05:57
There is simply no possible way black could have been played by a human ... just no way.  Actually this is exactly the type of game where a human would be extremely discouraged to see.  As white the human would have thought he had a dream position.  I am still not convinced that there is not more to white's game, I think that there is and most likely it would have to involve an earlier sacrifice of material.  Is it enough to win, probably not, but draw for sure yes!  Rybka took the bull by the horns and slammed the bull down, a very impressive game!
Parent - - By Maitra (**) [in] Date 2009-10-17 13:36
Hi- the first Rybka game White is Hermann -what program (I mean engine)is this?
Parent - - By Vempele (Silver) [fi] Date 2009-10-17 13:38

> Hi- the first Rybka game White is Hermann -what program (I mean engine)is this?


Doesn't your question answer itself?
Up Topic Rybka Support & Discussion / Rybka Discussion / 29th ODCCC 2009 in Leiden
1 2 3 Previous Next  

Powered by mwForum 2.27.4 © 1999-2012 Markus Wichitill