has been made with reversed engineering of Rybka. Well anyway
tried some new analysis. While less solid than Rybka, Fire(bird)
very fast sometimes finds interesting lines, and i seemed
to be able to maintain some slight advantage against both
the Zaitsev (9.. Bb7) and the Chigorin defences in the Ruy Lopez.
For example, whereas i previously thought black is ok in the
main line Zaitsev (9..Bb7 10.d4 Te8 11.Nbd2 Bf8 12.a4 h6 etc.),
with new analysis a rather obscure line, namely with 12.Ng5
seems to be able to achieve a white advantage.
After 12.Ng5 black plays 12.. Re7 to cover f7, and now
the (rare) move is 13.Ndf3! In Guez(2177)-Colin(2402),2005 black
played exd4?! (.. h6 also has been played, but 12.. Qe8N is
probably best), white played 14.cxd4 (Nxf7 would have been
better), and after 14. .. Qe8?! (Na5 would have been better)
we get the following position:
Now white could have got the advantage with 15.e5!N but
being a lower rated - and most likely worse- player,
he played 15.Bc2? and lost the game.
Well anyway the Zaitsev is anyway difficult to play,
especially for beginners, but the Chigorin still (9.. Na5)
is useful in practical play, because of the complexity
and many sideslines despite the slight advantage for white
according to my extensive analysis. For a similar reason,
black can also choose for other defences than the closed
defence, eg. 3.. g6 (Smyslov/fianchetto defence), etc.
See for example a recent book by Solokov on such lines.
But for advanced players, well now my analysis indeed points
to the Sicilian 1..c5 as 'best' defence for black.
Especially the lines with ..d6, possibly leading to
the Najdorf are solid, eg. 1.e4 c5 2.Nf3 d6 3.d4 cxd4
4.Nxd4 Nf6 5.Nc3 a6 but sometimes very sharp,
eg. with 6.Bc4!
And black also needs to know how to maintain equality
against 3.Bb5! instead of d4, with 3..Bd7 4.Bxd7 Nxd7
(or Qd7 c4!) etc. as described eg in a book by Pedersen.
One chessplayer advised me, that one should not
play the Sicilian below 2000, and he might be right,
a few weeks ago i played a youthful player who
in retrospect told me he wanted to play the Dragon.
But i simply tried the wing gambit, 4.b4!? after 3..d6
and won the game. So i'll keep on recommending
1..e5! in my book, except for advanced players.
and as for Fire(bird), well i expect the new version
Rybka 4 will be stronger again, especially if you
hire the online version, eg. on a cluster with 16 cpu's.
So more theoretical improvements expected to come
(update Dec 2010): with Rybka 4 the variation with Ng5
isnt best anymore, instead now against the Zaitsev
'best' play is going like 9.h3 Bb7 10.d4 Rfe8 11.Nbd2 Bf8
12.a3! with only a slight advantage for white.
Also with some new analysis in the Chigorin defence
against the RL (9..Na5) this variation appears to be
just as 'solid' as the Zaitsev, but also quite difficult.
(note: as result now the Petrov is advised for average
players with black, see the latest message in this blog
4 comments:
With Rybka, Fruit, and Firebird, have we reached a point in history where the average chess enthusiast can contribute to opening theory (and with little cost out of pocket)?
Amateur chess players are using a wiki to publish their analysis and to build on each others' analysis.
What are your thoughts? Will such a centaur chess wiki produce quality analysis?
Ref:
Final Theory of Chess Wiki Project
well i looked at this 'final theory of chess' project, but its incomplete, as the review on Amazon of the book also says, the author doesnt knows much about chess and is choosing an arbitrary repertoire; besides that, much analysis is done with Fritz, which i positionally not as good as Rybka.
While the idea of a wiki for chess theory is interesting, to do it properly one also needs a database to store the results, possibly minimax them to find the recommended lines, and/or add lots of (computer)games whereby statistics also can say a lot about
the ideal repertoire. Some earlier efforts (in database development) are used eg with the program Chess Assistant, using the C.A.P (chess analysis project) by D.Corbit. also here it depends on the engine with which the endnodes are analyzed, to determine which moves are the 'best'. But in fact this is rather arbitary, because opening theory is not only about the positional advantage which you can get with white, but also about the sharpness of the move. In other words , it matters a lot if you can reduce the nr of possible playing options for black and in such a way 1.d4 may be better than 1.e4. Newer engines in future might become better in such areas, but the result imho will be anyway that chess is a draw, and choosing a repertoire is a matter of personal choice, based on playing style. Using your repertoire in a accurate way, that is what matters, and not which repertoire (eg 1.c4, 1.Nf3 or whatever)..
Thanks for the reply and the info. As a fan of Wikipedia, I kinda think it’s a neat idea to try and develop chess theory in a wiki. I was just curious about what other chess players were thinking. I have some Rybka analysis of my own and might upload it to the wiki.
I saw the Amazon review but I chalked it up to the work of a troll after I saw these other reviews:
http://marshtowers.blogspot.com/2010/03/chess-reviews-136.html
http://gregschess.blogspot.com/2010/03/final-theory-of-chess.html
There is also a new review since I last checked.
http://chessbookreviews.blogspot.com/2010/04/gary-danelishens-final-theory-of-chess.html
Searching through the site I found a FAQ section which deals with some of the questions that I have had regarding the choice of computer program, etc. The following is a quote from the FAQ that gives a different twist:
Will the use of different chess engines by different contributors lead to a hodge-podge collection of inconsistent analysis?
Yes, the wiki will evolve into a composite of analysis produced by a variety of chess engines but herein lies the wiki’s strongest attribute for different chess perform better or worse depending upon the characteristics of the position (i.e. open vs. closed). The objective playing strength of the mainline continuations suggested on the wiki will, I believe, increase over time as contributors subject analysis already posted on the wiki to the scrutiny of a variety of chess programs. To paraphrase Linus' Law, "given enough eyeballs and chess engines, all chess problems are shallow."
The wiki will hopefully foster the creation of a collaborative social environment in which a decentralized peer-review process will emerge. Some contributors may be using the Fritz family of programs; others may use Rybka, etc. Of course, there will sometimes be discrepancies between Rybka’s top choice and Fritz’s top choice. In cases where chess engines disagree with each other, the community of wiki contributors will likely debate the merits of both variations. Hopefully, the objectively stronger move will win out in most cases and perhaps always win out in the long run.
The end result of this decentralized peer-review process will be the creation of a composite of the objectively strongest analysis produced by a variety of different programs, aided by human judgment. Through this process, the wiki may provide stronger analysis on a more consistent basis than any single chess program could on its own. After all, according to Kasparov,"Weak human + machine + better process was superior to a strong computer alone and, more remarkably, superior to a strong human + machine + inferior process."
That is a great tip especially to those fresh to the blogosphere. Simple but very accurate information… Thanks for sharing this one. A must read article! My web-site -
Teoriprov
Post a Comment