TN wrote on 04/03/10 at 09:46:36:

I had problem with 9.b4 variation (I played 9....Nh5 or Ne8 and f5 so Nf3 went for e6 etc ).
So I make a little research and I found some variations with 9...a5 that I liked it.
Maybe  black here must be well prepared.
So because of 9.b4 I look also the Na6 variation

I had problems with the 9. b4 variations so I give a try  at 6/7...Na6 variations also.
I liked for black the 9.b4 a5 .

BPaulsen wrote on 04/01/10 at 00:08:27:


Which huge thread are you talking about?

I'm not sure which I found more surprising, the idea that Black is nowhere near equality (perhaps on a par with the Old Indian at best, whereas I am used to seeing them considered equal or unclear with best play), or the idea that the only (or indeed the biggest) problem for White in the h3/Nf3 system is the ...c5 stuff.

I think there are a lot of top GM's that'd pay good money for your analysis of the Fianchetto and h3 lines if they show "nowhere near equality" for Black. Bologan seems to think that Black is equal in both places, and even Avrukh only found += in the Fianchetto, and who knows if that'll hold up.

Petrosian seems to be in the same boat as Korchnoi regarding black's concept in the KID.

On principle it probably `should` be unsound, but theory and practice are different issues.

I can see one of two things happening with the development of Classical KID theory going forward:

#1) White will get a major novelty somewhere in 9. Ne1 that finally gives him the sought-after +=, or failing that 7. Be3 will prove the improved way to go about this classical method of attacking black's set-up. 

However, if those don't work:

#2) White will start changing his approach to refuting the concept of black's set-up, with a more restraining/prophylactic system such as 6. h3, which in many ways attempts to work as an improved Petrosian System (6. Be2 e5 7. d5). This is noteworthy because the Petrosian System at one point caused black substantial problems, meaning that type of approach constriction approach (instead of chaos flank attacks) may ultimately be better.

Who knows...

I'm really torn by the thought that computers like the White side of the Classical KID.  I started playing the white side when I read Petrosian's comment that he thought the KID was fundamentally unsound.  (No, I don't know exactly where it was that I saw it.) This was long before computer analysis was reliable for more than 5 ply!

I do enjoy playing against the KID because it feels like I'm upholding the side of Right and Good, but I don't like supporting the side of an unthinking machine that can't appreciate Black's game.  The KID remains one of the great mysteries for computers, but I do tend to agree with them that White should be better! 

For one thing (speaking of Nakamura games), it is not at all the case that the soundness of the Mar del Plata depends on the "Give Mate or Die Trying" (as a recent NIC Yearbook article put it) approach associated with meeting 9. b4/Nd2 with 9...Ne8/Nd7.

kevinfat wrote on 03/31/10 at 19:16:29:


I'd encourage you to analyze it yourself. It's highly possible that Nakamura played it wrong and did indeed strike gold out of nowhere, but is it also possible that you hold too much faith in your bit-box? 

Are you guys sure about the Mar del Plata being sound. A lot of the positions in those lines (deep into games by Nakamura) look relatively simplified and computers really really like white's position.

That game is so fantastic! ...Qa8-a7! and Nh1!!-+

I don´t have no source, but but I read a couple of years ago, that Garri was quite sceptically about the correctness of the KID.
Afaik 9.Ne1 +10.Be3 is a killer in cc.

The responses are interesting. I've been looking at games in the Mar del Plata variation with Nakamura as black. Computers seem to absolutely love white's position near the end of the game until white makes some egregious blunder. Are computer evaluations bad to rely on here. I thought computer evaluations became more reliable further on into the game.

Net Warrior wrote on 03/31/10 at 13:55:16:


Black's "classical KID setup" in my conception is with the QN on d7 and pawn on c6.  The Mar del Plata is a system that I would be happy to play as Black based on my current understanding, but it's close enough to the edge that I don't think very many people would be surprised if a bust someday emerged, no, nor would it overturn anyone's fundamental conception of chess.  It's speculation, of course, since people have tried hard to bust it and failed so far. 

Now if the Slav were someday busted, that might be the same as saying Black is busted in chess, but I don't think that stands any chance of happening.

Top-level practice is in context of computers, of course, just not during the game.  And since this line is still relied upon at the top, it is indeed considered sound enough to play, based on everyone's current understanding.  The whole line is so complicated, though, that I don't think anyone would be terribly surprised if somebody came up with += or even a bust to Black's whole conception.  But a lot of people have tried, and it hasn't happened yet.

The kingside attack is sound at all levels, although you have to be well prepared as Black. This line is probably Black's best chance for equality, as Radjabov's and Nakamura's games have shown.