Schaakhamster wrote on 02/24/10 at 07:19:45:


I never said he was a weak player, but it's obvious when "analysis" is just the exact same thing as what an engine is giving/evaluating a position as.

Engines always think highly of IQPs because of the space advantage it confers, and the fact it's a weakness is usually beyond the horizon. Only when it becomes an actual target do they start to change their evaluation.

I think Sokolov's evaluation is what matters most here, given he has a better understanding of the position than any of us, or any engine would.

Ametanoitos wrote on 02/24/10 at 01:03:55:


I'm going to take an obvious guess - your "=" and "analysis" is based on computer evaluation and what it lists as its main line, isn't it? 21...Ng5 makes it obvious.

Computers will always overrate the chances of the side with an IQP. Always. They calculate, they do not take long-term positional factors like isolated pawns into account, and their evaluations do little to change GM perspective (in this case Sokolov's). They will only change to += much later when it becomes clear that the isolated pawn is an actual target.

There's a lot of Tarrasch Defense positions that theory evaluates as += that engines mark as equal, and it can also be seen in other IQP positions (ie: those arising from the QID, and 5. Bf4 QGD).

We will deal with your sideline later Mnb! For now:

After Sokolov's 18.h3! i like 18...Ne5 better than 18...Ne4 but 18...Ne4 is typical when h3 has been played. I dont understand why you say it does nothing. It is very typical: closes Bg2 diagonal, sometimes Ng5 is a threat (or Nxg3/xf2) when the weakness of he shows f.e 19.Qe2 Bf5 20.Nb5 Ng5!? 21.Kh2 (21.h4?! is obviously weakening) 21...Be4 =. Also 19.Re2 Nb4! seems strong with an ideal of exchange on c3, but there are other 19nth moves for Black.

Also after 14.Bd4 i think Black should not take on d4 and play 14...Ne4 (14...Rc8 also looks OK because 15.Nc5? Nxd4 is the idea. Maybe after 15.e3! b6 is OK with the idea Qd7)
15.e3 Bb4! 16.Na4! Qe7! 17.a3 Bd6 18.Nac5 Bf5 19.Nd3!? Rad8 20.Nf4 Bb8 and i prefer Black here. Ofcourse this doesn't mean anything just that i like Black's chances but i am a "dwarf" also!  So...

I have put the 9 Bg5 cxd4 ...... 11 Bg4 into pgn compiling all the analysis thus far. If I have time, I will compile for the other variations as well.

Ametanoitos wrote on 02/23/10 at 11:33:59:


So... let’s see what “dwarf” Keilhack has on 11...Bg4. 
First he is a bit reserved, calls it a move that’s quite common in the middle ranks, but to quite a large extend mistrusted in the GM ranks. "It’s attraction may come from that it looks more active than 11...Re8, but it is a fictuous activity". Keilhack’s main line is 12.Qa4. 

With the variation you quote from Sokolov you run in a damned problem that blossoms in the whole 9...cxd4 main line: Move ordering. With 12.Nb3 Be6 13.Rc1 Re8 you have drifted into 11...Re8 12.Rc1 Bg4 13.Nb3 Be6.
Doesn’t matter, you might say refering to your findings after 14.Re1 Qd7 15. Bc5 Rac8 (15...Rad8 - Geller) 16. Bxe7 Qxe7 17. e3 Red8 18. h3!N (Keilhack only has 18.Qe2 Bg4! 19.f3 Petrosian - Spasski wc m 4 1969, with a severe battle to follow and gives 19.Qf1!?) 
"...securing White a small advantage", Sokolov. He is right I feel. Black’s bishop is bit bad. Your 18...Ne4 isn’t connected to really active options (but I only had it on the board and tend to overlook deeper tactics), white has smothered some of your possibilities with h2-h3. Because black has no dark bishop, the slight weakening of white’s king’s fortress, which in other variations (with dark bishop on board) gives attacking possibilities, is not an unsurmountable problem if he is aware of the sting against h3 (e.g. with Qd7 in combination with Ne4's kick on f2). So white shouldn’t hurry. On 19.Qe2 you give 19...Bf5 20.Red1 but that fails to keep black at bay as you show. What about say 20.Nb5 overprotecting d4? And I’m toying with 19.Re2, protecting f2 with a glance on c2 too, perhaps later again with sth. like Nb5.

And you have to deal with the interesting move 14.Bd4!? Keilhack gives with some lines. On 14...Nxd4 15.Qxd4 the queen is centralized and white plans Nc5 and/or Rfd1. The bishop as a blockader is unusual but it is difficult to handle for black who is tied to d5 and has to keep an eye on options like Bxf6 chopping down a defender or white manoeuvres behind the bishop.


But as said Keilhack’s main line is 12.Qa4 right in Sokolov’s way 12...Na5 13.Rad1 Nc4 14. Bc1. He calls 14...Nb6 15.Qb3 as obviously passive (and gives - with deviations and lines - 14...Qc8 15.Qb5! Nb6 16.Bf4! asf. in the end resulting a white plus in Stull - Poulsen, corr. 1989).
Keilhack also quotes that game Vadasz-Szilagyi, Budapest 1974, but calls the course of the game not really evidential 15...Qd7 16 Be3 Bc5?! 17.Nxe6! as Khalifman (or he as him...). But Keilhack writes on 15...Qd7 “seems logical, but it is not good due to the shaky position of the Nb6." Without words and lines he gives 15...Rc8 as the alternative As I saw in your last post Sokolov has worked on this idea with some lines.
Black’s mainline against 12.Qa4 is acc. Keilhack 12...Qd7 which offers d5 (and if white takes - best with 13.Bxd5 - the conclusion is that black is worse but only in relatively slight dimensions). After 13.Rfd1 an interesting game with tactical possibilities rises.
There are structural or theoretical defects in black’s position but there are tactical or practical chances too.
GM Marjanovic is the master of ceremony in the 11...Bg4-variation and played many games with it. So there is a strong player who constantly believes in it for many years. An argument, that should count. 
Keilhack’s conclusion is: "Whether the often proved richness of possibilities compensates for the theoretical soft spots of this variation, every player has to decide on his own."

I think, that is overall the very core of the Tarrasch...
cheese

Computers are notorious for overrating the chances of any side that has an IQP. Using them to check moves is fine, but for evaluations? Not so much. Computers base their evaluations on what they can calculate immediately, long term positional weaknesses don't matter unless they can be exploited in their calculations. Horizon effects at its finest...

Sokolov's evaluation has the most consequence.

Schaakhamster wrote on 02/23/10 at 05:34:57:


Yes, Keilhack works on it. More than a third of page... 
To my mind it makes no difference whether you only suggest a move or play it. The first is called theoretician the latter practician. So or so it's food for thought and a good food at that, else the better players wouldn't eat it.
It's not allways us standing on the shoulders of giants. Sometimes that giants stand on the shoulders of us dwarfs.
I confer to the story of a gambit in KID 4-pawns-attack. 1.d4 Nf6 2.c4 g6 3.d4 Bg7 4.e4 d6 5.f4 0-0 6.Nf3 c5 7.d5 e6 8.Be2 exd5 9.e5?! which for some time troubled black (til 9...Ne4 was found to be a good remedy -- AND please don't discuss on this here... it's only a very illustrative example).
It was invented (to my ressources at least), analysized and  largely advocated by the amateur and resourceful theoretician Gerhard Gunderam, who dedicated some chapters to it in his even today inspiring 1961 book "Neue Eröffnungswege I" and too 1967 in "Neue Eröffnungswege Folge 2".
He comments "8...exd5 That's well known and nearly self-evident... 9.e5 ...but not that. A sacrifice to stimulate the play".
He didn't shy away from molesting wellknown grandmasters like Keres and even ex-WC Euwe with letters on his findings. If I remember that correctly, that gambit is sometimes called Gunderam-Gambit and it was taken up by GMs as databases may show.

And Gunderam or Keilhack by far aren't the only examples of mindfull "dwarfs". Even I find new paths in GM-play from time to time...

@GabrielGale: Sorry for being suspicious. I think I am a lesser twin of Stefan Bücker (and my scientific inclination breaks all to often through). He allways instists to quote precisely the resourses - and when working on a topic to check the important books earlier on published. As it seems Cox wasn't aware of Keilhack...

cheese