Not logged inRybka Chess Community Forum
Up Topic Rybka Support & Discussion / Rybka Discussion / Game of note for Rybka opening book
- - By FirebrandX (**) Date 2007-12-28 11:37
In a recent engine versus engine CC match, I used the "Rybka II" opening book and played the black side of a Queens Indian. I followed the book's tournament choices on each move, but quickly got into what was probably a losing position after VERY strong play from white starting from move 25. In the Rybka II book, move 25 has Qc2 shown for white, which gives black excellent chances for the win. My opponent played Qc1 instead and initiated what was probably a winning assult on my kingside. I only managed to escape with the draw by doing some deep forward-sliding to trick my opponent's engine into a drawn line that starts off with a winning eval for white (I consider this type of draw to be almost as good as winning). Here's the game:

1. d4 Nf6 2. c4 e6 3. Nf3 b6 4. g3 Ba6 5. b3 Bb4+ 6. Bd2 Be7 7. Bg2 c6
8. Bc3 d5 9. Ne5 Nfd7 10. Nxd7 Nxd7 11. Nd2 O-O 12. O-O Rc8 13. e4 b5 14.
Re1 bxc4 15. bxc4 dxc4 16. Qa4 Bb5 17. Qc2 Re8 18. a4 Ba6 19. Rad1 Bf8
20. Nf1 Rb8 21. Ne3 Bb4 22. e5 Qc7 23. f4 Bxc3 24. Qxc3 Rb3 25. Qc1 Nb6
26. f5 Kh8 27. Be4 Nd5 28. Nxd5 cxd5 29. Bc2 Rbb8 30. f6 Rg8 31. Rf1 gxf6
32. Rxf6 Rg7 33. Rdf1 c3 34. Rf1f2 Bc4 35. Qa3 Ba6 36. Kh1 Kg8 37. h3 Qa5
38. Qc1 Bc4 39. Rh6 Qd8 40. Rff6 Qf8 41. Rf4 Be2 42. Qe3 Qa3 43. Rxh7 Bf3+
44. Qxf3 Qc1+ 45. Qf1 Qxf1+ 46. Rxf1 Rxh7 47. Bxh7+ Kxh7 48. Rxf7+ Kg6
49. Rc7 Rb3 50. Kg2 Kf5 51. Kf3 c2+ 52. Kg2 Rb2 53. Kf3 Rb3+ 54. Kg2 Rb2
55. Kf3 Rb3+  1/2-1/2
Parent - - By Nelson Hernandez (Gold) Date 2007-12-28 12:28
You might have won.  Look at 25...c3.  78% in my book, with apparently winning lines among all the likely replies.
Parent - - By FirebrandX (**) Date 2007-12-28 12:43
The Rybka II book had no data on Qc1, so I was left to my own devices. I'll have a look at c3. BTW, the c3-pawn did end up saving me from the loss. Its probably the most important point of the entire opening line when you reach the 25. move here.

If I may inquire, what book are you using?
Parent - - By Nelson Hernandez (Gold) Date 2007-12-28 12:54
It's a private creation that a couple of crazy guys haphazardly slapped together.
Parent - - By FirebrandX (**) Date 2007-12-28 13:16
Ah, I understand. ;)

Just one more quick question if I may,

What is considered to be the best public book for engine-engine use? From what I gather, private books are more powerful, but what would be the best bet for use with Rybka 2.3.2a that I can get a hold of?
Parent - - By Nelson Hernandez (Gold) Date 2007-12-28 15:25
You have asked a trick question.  I am now in a kind of zugzwang: anything I say on this subject could result in my making unwanted enemies.

The truth is I don't know and I doubt anyone else really knows.  First, nobody has the time and tools to rigorously test a set of books measuring the right things.  Further, any answer on any particular book would not be constant if measured against performance in competition; public books all tend to weaken over time. 

You should avoid books that bill themselves as "neutral" as those are best used for engine testing.  (No book is really neutral.  But those that seek neutrality can provide common benchmarks.)  You should also avoid books that have been around longer than 6-12 months, as most likely they've been played to death and people have exploited their weaknesses.  Within those constraints I am inclined to favor Harry Schnapp's books provided they are fairly recent.  He seems to be a master of sharp lines.
Parent - - By PAKman (**) Date 2007-12-28 15:54
Well stated Nelson.
Parent - - By Nelson Hernandez (Gold) Date 2007-12-28 16:29
Grateful I didn't make a new enemy out of you, Norm!
Parent - - By FirebrandX (**) Date 2007-12-28 17:33
Thanks for the tip!

Athough I'm not sure which is the better book to use... The masterbook or the slightly older rybka book?
Parent - - By Lukas Cimiotti (Bronze) Date 2007-12-28 18:14
Nelson is perfectly right with what he says. Btw he's one of the best book makers, but he won't tell you ;) And he won't give you his book. But books get outdated much faster, at least if you play on Playchess.
Both HS books are really good (I prefer the masterbook), Sheebar is very good, too. For long games, still RybkaII book is very good.

Parent - - By Nelson Hernandez (Gold) Date 2007-12-28 23:27
I am very grateful to be recognized as "one of the best", especially considering that I don't think anyone has said that before.  There are so many comments I could make to debunk this assessment, however!

Nonetheless thanks for your kind words.
Parent - By Roland Rösler (****) Date 2007-12-29 03:20 Edited 2007-12-29 03:24
Who is "one of the best" in cooking opening books? I think, there are some misunderstandings and unclarities. For me, the best book is the book, when you can just play chess after the opening (bothsides :-)). Therefore I love Gata Kamskys remark in recent interview: Well, the most important thing is to prepare good openings (laughs). And then I can just play chess.
And the Cato team always play chess in freestyle tournaments :-). Because of the opening book of Nelson? We don´t know! We only see, that Anson and Wing never have problems with the opening (similar to team Rajlich, Klosterfrau, Rentner2 or Ciron).
PS: Nelson, keep cool! This was the best compliment, I can make. I have no clou about openings. But I see the games ...
Parent - - By Henrik Dinesen (***) Date 2007-12-31 12:04
Well Nelson...

I have no reason to recognize you as a one of the best... And no reason no to :-)
Still having a somewhat bad feeling since I "bailed" out on our game-trade 3 years ago...

I know you do things very well, and want to wish you a happy new year, live long and in prosper :D

Parent - By Nelson Hernandez (Gold) Date 2007-12-31 17:21
Heh, I had totally forgotten that.  Forgive and forget, that's what I say.  But if your feelings of regret extend to now wanting to send me your entire game collection I promise I will put it to good use against the rest of these mugs!
Parent - - By Sesse (****) Date 2007-12-29 00:37
Asking what book is the best feels a bit like asking what the best opening move is. In general, any public, popular book will eventually weaken (as Nelson pointed out); others will spend time coming up with refutations of the lines in there, or almost-as-good lines that are not in the book to throw you out of it and then eventually lead the engine into traps.

...or so I've gathered. I've never played public engine-engine chess, nor made opening books. :-)

/* Steinar */
Parent - By FirebrandX (**) Date 2007-12-29 02:18
Thanks for the repeat of what Nelson posted. I fully understand this, however, I was merely asking for the strongest public book. Yeah yeah, I know they are weakened by being public books, but I was merely after the best I can use compared to other public books. Understand where I'm coming from now?
Parent - - By Nick (*****) Date 2007-12-29 00:58
Yeah 25 ... c3 seems decent ...

[Event "Partita con Elo, 3m + 0s"]
[Site "Gioco con programmi"]
[Date "2005.03.29"]
[Round "?"]
[White "Cato the Younger, Shredder 9"]
[Black "Auryn, Shredder 9"]
[Result "0-1"]
[ECO "E15"]
[WhiteElo "2664"]
[BlackElo "2592"]
[PlyCount "136"]
[EventDate "2005.12.30"]

1. d4 Nf6 2. c4 e6 3. Nf3 b6 4. g3 Ba6 5. b3 Bb4+ 6. Bd2 Be7 7. Bg2 c6 8. Bc3
d5 9. Ne5 Nfd7 10. Nxd7 Nxd7 11. Nd2 O-O 12. O-O Rc8 13. e4 b5 14. Re1 bxc4 15.
bxc4 dxc4 16. Qa4 Bb5 17. Qc2 Re8 18. a4 Ba6 19. Rad1 Bf8 20. Nf1 Rb8 21. Ne3
Bb4 22. e5 Qc7 23. f4 Bxc3 24. Qxc3 Rb3 25. Qc1 c3 26. Qc2 Reb8 27. Rc1 R8b4
28. f5 Qb8 29. Qe4 Qc7 30. fxe6 fxe6 31. Bf1 Nc5 32. Qg4 Qc8 33. Bxa6 Nxa6 34.
Rf1 Nc7 35. Qg5 Nd5 36. Qh5 Qd7 37. Qf3 Qe8 38. Nxd5 cxd5 39. Rxc3 Rxc3 40.
Qxc3 Rc4 41. Qb2 Rxa4 42. Rf2 Rc4 43. Rf3 a5 44. Kg2 a4 45. Rf4 h6 46. h4 Kh7
47. Rg4 Qh5 48. Qb1+ Qf5 49. Qxf5+ exf5 50. Rf4 Kg6 51. g4 fxg4 52. e6 Rc2+ 53.
Kf1 a3 54. e7 Rc8 55. Rf2 Re8 56. Ra2 Kh5 57. Rxa3 Rxe7 58. Ra5 Rd7 59. Kg2
Kxh4 60. Ra1 h5 61. Ra5 Kg5 62. Rb5 Kf4 63. Rb8 Ke4 64. Kg3 Kxd4 65. Kh4 Ke3
66. Re8+ Kf2 67. Re5 g6 68. Rxd5 g5+ 0-1

Parent - - By Nelson Hernandez (Gold) Date 2007-12-29 13:05
Hoist by my own petard!  You clever rascal.  Well, bear in mind this was played on a P4 back when my opening book was a wee lad. 
Parent - - By Banned for Life (Gold) Date 2007-12-30 00:22
Nah, your opening book was already top shelf by December 05. As I recall, you were a member of a small group with slow computers whose members were still a threat to win major engine room tournaments against much faster duals and quads (Major in this case meaning a prize fund of over $50!!! :-) :-) :-) ).

Parent - - By Nelson Hernandez (Gold) Date 2007-12-30 00:54
Yeah, that's true.  Hmm, and that was two years ago.  I don't suppose we've stood still, either.  But vulnerabilities remain: that terrible loss to Eros still resonates.  My god, he just dumped the whole pasta salad on our heads.  Freakin' guy imposed a Carthaginian Peace on us.
Parent - - By Banned for Life (Gold) Date 2007-12-30 07:01
The Eros freestyle fiasco seems kind of strange to me. You both have large databases and I would assume there is a great deal of overlap between them. I don't think Eros does much original development, so I'm assuming he came across and squirreled away the trap line you fell into, but your team should be the least likely to be susceptible to this kind of thing.

Parent - - By Nelson Hernandez (Gold) Date 2007-12-31 17:25
We played French against him; in Anson's defense, it was my idea.  What we've concluded is that French is more apt to produce positions that favor chess knowledge of the GM variety, not exactly our greatest strength.  Not saying we'll never play it again, but surely we were not ready to play it that day against that opponent.  We knew we were in serious trouble from about the sixth move.
Parent - - By Banned for Life (Gold) Date 2007-12-31 22:47
Fair enough, but given your possession of a large database with both win-lose-draw statistics and engine evaluations, I would think that against a DB statistics player like Eros, you would look for positions with divergent viewpoints and find the ones where the statistics were misleading and the engine values are correct. This is the opposite of what I normally do, which is trying to end up in divergent positions where my private database is correct and the engine evaluation is wrong (which is most likely to occur in the first 10 moves of the game since most engines have a harder time evaluating opening positions than middle game positions).

Parent - By Nelson Hernandez (Gold) Date 2008-01-01 02:25
We don't systematically use either approach you describe, exactly, but our actual play very often (insensibly) amounts to the same thing.  One of the toughest things in making move decisions in the opening is properly balancing empiricism with evaluation.  We've worked a lot on this in the last several months.  It's not easy.  There are no universal solutions.  At best there are formulaic approximations that statistics show work pretty well most of the time.

Not to diminish Eros' achievement, but we had multiple games in play that round: two tough opponents at the same time.  We just weren't all there against him in terms of attention or hardware.  Not claiming the outcome might have been different, but you never know.  The only thing I know for sure is that we were both shaken and stirred.
Parent - - By Dhanish (***) Date 2008-01-01 08:03
Thankyou for the game. Is there a way to download or search for such Engine vs Engine games? This line did not appear in all the databases that I searched!
Parent - - By PAKman (**) Date 2008-01-01 12:41

There are some 2007 Engine vs Engine games at

Happy New Year

Parent - - By Dhanish (***) Date 2008-01-03 05:39
Hi Pakman,

Thanks for the link. An interesting site. The playchess ER file links do not appear to be working at the moment. Perhaps you could include the number of games / file size near the link in the web page itself.

I am glad you realise that there is a need to make engine games available for download. Unfortunately, these games are from 2007 only. There are plenty of sites for human games, but not many for engine games. Perhaps somebody who has collected games earlier could make them available somewhere. Those very long back, when engines were weak, and those on very short time controls, may perhaps be excluded.

Regards, Dhanish
Parent - By PAKman (**) Date 2008-01-03 23:48

The Playchess Engine Room games are in Chessbase .cbv format.
I have limited space available on my website.
I have 540,000 playchess games from 2006, but not enough room on the site.
I am working on this problem, and hope to have it solved in a couple weeks.

Thanks for your interest.
Up Topic Rybka Support & Discussion / Rybka Discussion / Game of note for Rybka opening book

Powered by mwForum 2.27.4 © 1999-2012 Markus Wichitill