Reports | March 30, 2007 19:55

[lang_nl]Een belangrijke verbetering?[/lang_nl][lang_en]An important improvement?[/lang_en]

[lang_nl]Een opening die de laatste tijd op topniveau niet erg populair is, is de Tarrasch Verdediging. Je ziet vooral Slavisch, Dame-Indisch en dankzij Radjabov tegenwoordig ook weer veel Konings-Indisch. Hoewel Alexander Grischuk de Tarrasch soms als gelegenheidswapen gebruikt, lijkt er sinds Kasparov in de jaren tachtig geen topspeler meer te zijn geweest die de opening serieus met zwart bestudeerd heeft. De reden hiervoor is misschien dat veel zwartspelers de Tarrasch onbewust associeren met saaie stellingen en maximaal remise als resultaat. Dit is een misvatting.[/lang_nl][lang_en]An opening that is currently not very popular at the highest level, is the Tarrasch Defence. You mainly see Slavs, Queen's Indians and thanks to Radjabov many King's Indians as well these days. Although Grischuk uses the Tarrasch occasionally, no top player seems to have seriously studied the opening since Kasparov in the eighties. The reason might be that Black players unconsciously associate it with boring positions and a draw as the best possible result. This is a misconception.[/lang_en]

[lang_nl]De Tarrasch is een dynamische opening waarin zwart actief stukkenspel krijgt in ruil voor een geisoleerde d-pion of een zwakke c-pion, die slechts tot remise leidt als wit geen winstpogingen doet of niet doorstastend handelt. In mijn eigen zwarte praktijkervaring heb ik met de Tarrasch Verdeding een ruime plusscore tegen gelijke ratingtegenstand. Tegen niemand minder dan Loek van Wely maakte ik er in een snelschaakpartij onlangs nog remise mee. Ook in de Meesterklasse blijken witspelers grote moeite te hebben om 'erdoor heen te komen', zelfs als daar specifieke voorbereiding aan te pas gekomen is.

Vorige week kreeg ik weer een Tarrasch op het bord in de wedstrijd van mijn team Homburg Apeldoorn tegen LSG. In die partij kreeg ik de gelegenheid een verbetering op een aanbeveling van Khalifman te spelen, die naar mijn bescheiden mening de angel haalt uit ?ɬ©?ɬ©n van de belangrijkste bestrijdingswijzen voor wit.

>> open in een nieuw venster

Als de zet 16...Dd7! (of het hele idee De7-d7) inderdaad ook na 16.Da4 een verbetering is op de bestaande theorie, betekent dat volgens mij dat wit op zoek moet naar een andere manier om voordeel te behalen tegen de Tarrasch. Misschien met de 'positionele' benadering 9.dxc5, of toch met de scherpe zet 11.f4!? Waar ik vooral benieuwd naar ben is waarom Khalifman, die het idee wel kende na 15.Dc2, de zet Dd7 in deze stelling niet noemt.

Todat ik antwoord op die vraag krijg, is mijn conclusie dat zwart leeft in (deze variant van) de Tarrasch...[/lang_nl]
[lang_en]The Tarrasch Defence is a dynamic opening in which Black obtains active piece play in exchange for an isolated d-pawn or a weakish c-pawn, which only leads to a draw if White refuses to try and win or doesn't act pointedly. In my own experience with Black, I have a pleasant overscore against equal rated opponents. In a recent blitz game against Loek van Wely, I managed to make an easy draw with it. In the Premier League of the Dutch competition, my White opponents also seem to have trouble 'breaking through', even if they prepared specifically for it.

Last week, in the match of my team Homburg Apeldoorn versus LSG from Leiden, I got a Tarrasch on the board again. In that game, I got the opportunity to play an improvement on a Khalifman recommendation, which, in my humble opinion, solves the problems in one of the most important lines for White.

>> open in a new window

If the move 16...Qd7! (or the entire idea Qe7-d7) after 16.Qa4 is indeed an improvement on known theory, this means in my view that White has to start looking for another way to obtain an advantage against the Tarrasch. Perhaps with the 'positional' approach 9.dxc5, or with the sharp move 11.f4!? The question that intrigues me most is why Khalifman, who did know the idea after 15.Qc2, doesn't mention the move Qd7 in this position.

Until I get an answer to this question, my conclusion is that Black is alive in (this variation of) the Tarrasch Defence...[/lang_en]

Arne Moll's picture
Author: Arne Moll


Eduardo's picture

Arne: Perhaps I am being too dogmatic. It would be interesting to check the databases to see how Black scores with 9...c4 and then compare it with 9...h6 followed by cd4. I have a gut feeling that 9...h6 outscores 9...c4.

arne's picture

There is actually a great monograph on the Tarrasch Defence, by Harald Keilhack. See
for some details. Khalifman, too, deals with the Tarrasch in an excellent way. I somehow have the feeling both books are much better and deeper than the two books you're mentioning, even though I haven't read them myself.

arne's picture

@Eduardo: I think it's too dogmatic to say that there 'must' be something wrong with Black's position after 9...c4. Why is that a 'must'? The fact that new ideas are born doesn't mean they have to be wrong. It's a matter of choice; of course taking on d4 is also interesting but it's a different game. Surely c5-c4 is a very principled decision in itself, definitely worth investigating.
I also don't agree that White is playing for two results after that. If White wants a draw, he can surely get it with correct play, but nobody is perfect and if he wants more, he also has to take risks. It's not a free ride, as practice shows.

Eduardo's picture

I beg to disagree with you. I used to play the Tarrasch too. If you have to play 9...c4 in order to equalize than something must be wrong by definition as the idea in the Tarrasch is to have good piece play in exchange for the isolated pawn as you pointed out. Look at the games of Kasparov during the 80s, he never played 9...c4. After 9...c4 10.Ne5 White is playing for two results and Black only for one. The more principled move (as the Eastern European GMs like to say) is 9...h6 followed by cd4. If am not wrong theory currently gives White an edge in this line too but Black has lots of alternatives to complicate the game (thus increasing his probabilities to play for two results instead of just one as in the line with 9...c4).

Killiedoc's picture

Don't think I like white's game with queen stuck on the a file, however some interesting ideas put forward. Many thanks!

Eduardo's picture

I offer you what I think is a key critical game in the Tarrasch Defense. White won convincingly but perhaps there has been an improvement for Black since the time this game was played.

[Event "Kochin Wch-jr U20"]
[Site "?"]
[Date "2004.??.??"]
[Round "5"]
[White "Harikrishna, Pentyala"]
[Black "Petrosian, Tigran L"]
[Result "1-0"]
[ECO "D34"]
[WhiteElo "2612"]
[BlackElo "2539"]
[Annotator "M/05-1-83 Harikrishna"]
[PlyCount "119"]
[EventDate "2004.??.??"]

1. d4 d5 2. c4 e6 3. Nc3 c5 4. cxd5 exd5 5. Nf3 Nc6 6. g3 Nf6 7. Bg2 Be7 8. O-O
O-O 9. Bg5 cxd4 10. Nxd4 h6 11. Be3 Re8 12. Qb3 Na5 13. Qc2 Bg4 14. Nf5 Bb4 15.
Bd4 Bxc3 16. Bxc3 Rxe2 17. Qd3 Re8 18. Ne3 Be6 19. Qb5 b6 20. b4 Nb7 21. Bxf6
Qxf6 22. Nxd5 Bxd5 23. Qxd5 Re7 24. Rae1 Rc7 25. Rc1 Rac8 26. Rxc7 Rxc7 27. Rd1
g6 28. h4 h5 29. Rd2 Qc3 30. Be4 Kf8 31. Qg5 Kg7 32. Kg2 Qf6 33. Qd5 Re7 34.
Bf3 Rc7 35. Be2 Qe6 36. Qxe6 fxe6 37. Bf3 Kf6 38. Be4 a5 39. a3 axb4 40. axb4
g5 41. hxg5+ Kxg5 42. Kh3 Kf6 43. f4 Ke7 44. Bxb7 Rxb7 45. Re2 Rc7 46. f5 Kf6
47. fxe6 Ke7 48. Re5 Rc6 49. Kh4 Rc4+ 50. Kxh5 Rxb4 51. g4 Rb1 52. g5 Rh1+ 53.
Kg6 Rb1 54. Kh6 b5 55. g6 Rh1+ 56. Rh5 Rg1 57. Kh7 Kf6 58. e7 Re1 59. g7 Rxe7
60. Rh6+ 1-0

Eduardo's picture

Arne: Your reference database apparently supports your claim. However I believe the statistics you present are biased because there is a large number of games from weak players in the line with 9...h6 or 9...cd4. Should you generate a database only with games of strong players (say 2400+) I bet the statistics would tell a different story. Just as a reference Gary Kasparov played this position with Black after 9.Bg5 11 times against high level oposition during the 1980s and never, ever, played 9...c4. He always went for 9...cd4 or 9...h6. That tells me the greatest player in history does not think highly of 9...c4.

arne's picture

@Eduardo: a few things come to mind.

First of all, I would like to remind you that I was the one who said statistics didn't mean so much. For argument's sake, I still decided to put in some rough statistics, but only to support this general statement. Now, however, you seem to wish more detailed statistics from me in order to support your own hypothesis - not mine - for which by the way you have given no concrete evidenceor statistics in the first place, except the claim that Kasparov has never played it.

Well, let's look at this claim first. Now, Kasparov has probably never played most of the opening lines in chess theory (for example, he has probably never played the French Defence or the Caro-Kann) but does that say anything about the soundness of them? I don't think so. Fischer never played 1.d4, but surely that doesn't mean that Fischer didn't trust this move or that he thought it was unsound. You also have to think about personal taste, about available time to study new opening lines, about opening fashion, about specific preparation for opponents, about available books and study material, about surprise weapons and about a good knowledge of your own strenghts and weaknesses in certain positions. Perhaps Kasparov felt that his strenghts could not be at their best after 9...c4. Perhaps he really didn't like Black's position after 9...c4 Kasparov, somewhat notoriously, also didn't like Black's position after 1.e4 e5. But it really doesn't tell us anything about the objective value of the Open Games.

But even if we do accept the idea that Kasparov didn't trust the 9...c4 line, I would still like to point out that this was in the early 80s and we're living now in 2007. Chess opening theory has developed hugely in that time. Even Kasparov was not 25 years ahead of his time. And ideas about what 'must' be good or bad in openings have changed enormously in the past. For example, just take a look at 'Secrets of Grandmaster Strategy - advances since Nimzowitsch' by John Watson. It should give you a good idea of what sort of things have changed in our perception and understanding of (opening) chess theory. I would like to give just one example. Only 20 years ago, most GMs were of the opinion that the Sveshnikov line of the Sicilian was completely unplayable for Black, and that it consisted of a positional mistake to allow the weakness on d5. But now everybody plays it and nobody worries anymore about the so-called weakness on d5. It's just an interesting middle game position, and let's not be worried about the endgame just yet. Fischer said the Dragon was a win for White, but still Kasparov himself played it much later, and even in World Championship match. (I think it was against Anand.)

Also, to backup my claim that statistics don't mean much, let's consider the fact that one important novelty may change a variation or even an opening completely. Thus, even though for decades White has scored hugely in some line, after some important novelty or (series of novelties) by Black, this picture may change completely. But the statistics will still show a huge statistical advantage for White for years and years to come. But surely, this doesn't mean much anymore. In fact, the point I was trying to make in my article was that perhaps this Qd7 move could transform the look of the 9...c4 line, which by the way wasn't so bad in the first place, even according to Khalifman. Such things have happened numerous times.

Besides, I think most of your argument is a bit circular reasoning. On the one hand you seem to 'bet' (but doesn't support with concrete statistics) that strong players tend to score better with cxd4 and h6 (2400+). Well, perhaps. We could analyse it. But on the other hand, you say you suspect that there is a large number of games of weak players with 9...cxd4. Now, do you suggest that 9...c4 is NOT played so much by weak players, but only by strong players? This should affect the average rating, but it clearly doesn't, as my first statistics show.
But even if it did, it would surely mean that strong players prefered 9...c4 and this would, in fact, be also contrary to your argument. So it seems something is wrong there.

Finally, I just did a quick seach on my database - again just for argument's sake :-) - on the 2500+ statistics (both players having 2500+ rating) of both 9...cxd4 and 9...c4 lines. It turns out in both lines Black scores exactly 39%! Well, like I said, in my opinion this doesn't say much about the general value of both lines, but perhaps it can convince you that really, both lines are perfectly playable?

Regards, Arne

Tom Chivers's picture

Are you sure Bert? I remember reading an interview with him where he said he played it once; his coach told him not to be a coward, so he switched to the Sicilian... Or at least something along those lines?

arne's picture

@Tom: just look in the databases, he played it many times until 1982.
In the book 'Kasparov's Fighting Chess' there is the following comment by Speelman on the game Eolyan-Kasparov, Riga 1977: '1.e4 c6 Gary frequently adopts variations from the opening repertoire of his trainer, Botvinnik, who used the Caro Kann so effectively in his 1958 world championship match with Smyslov and his 1961 match with Tal'.
So, it looks like the story about the coach is just an urban legend.

yyyacb's picture

"Why are there no opening monographs on the Tarrasch Defense?"

On this topic, here's a link to my blog entry.

arne's picture

@Eduardo. Well, average scores don't say everything, but I will accept the challenge. In my Chessbase Reference database (probably not up to date, but good enough for some significance), Black scores 39% with 9...c4 with an average rating of 2338, and 38% with 9...cxd4 with an average of 2334. 9...h6 scores much worse: only 19% with an average of 2190. So, judging on these figures, I'd say 9...c4 is at least as sound as 9...cxd4.
Moreover, I think 9...c4 is actually more principled that 9...cxd4. After all, Black can reason that White has missed the opportunity to give Black an isolated pawn on d5 with 9.dxc5 and is now profiting from this positional mistake. To be sure, this reasoning is a bit too optimstic, but perhaps it shows why top players like Spassky, Short and Grishuk have prefered it over 9...cxd4

Bert de Bruut's picture

Arne, the Caro-Kann was the mainstay of Kasparov's repertoire vs 1. e4 when he was young...

arne's picture

Thanks, Bert. I didn't know that! Originally, I just wanted to mention the French, but I suddenly realised I had never seen him play Caro-Kann either. Didn't bother to check, of course. It seems he played his last serious Caro-Kann as Black in 1982. Well, I was eight at the time, so you'll forgive me I'm sure ;-)(Actually, it just proves my point. The fact that Kasparov stopped playing it at some point, certainly doesn't mean the opening is not playable anymore...)

Eduardo's picture

Hello Arne! Calm down, this is just a friendly chess chat! Obviously we have contrasting views regarding this particular line of the Tarrasch. You make however a good point when you make reference to the Sicilian Sveschnikov. It may very well be that this Tarrasch line with 9...c4 is one of these opening moves that look bad but end up working after all due to certain concrete nuances of the position. Time will tell. I stick with Kasparov's approach and believe that 9...c4 is inferior to the plan with cd4 and h6. Keep up with the good work and congratulations for this excelent chess blog!

Latest articles