Choices and doubts (part I)
3 January 2007 5:32 PM | Last modified: 23:22
The cool thing about analysing your own games is that you can never tell beforehand how long you will be analysing. A serious tournament game against a strong opponent might be forgotten only hours after the game, while a relatively unimportant club game can haunt you for weeks. Recently, I played such a game on the competition of chessclub Max Euwe Amsterdam.
The game contained a couple of moments in which both players had to make tough, game-deciding choices. In such situations, a little grip would be quite pleasant, but unfortunately we didn’t have this grip, or not enough anyway. In this first part I will show the first important moment and the doubts both players were having, and illustrate it with some games from top players.
Moll-Hoffman
Amsterdam 2006
This is the position after Black’s 13th move, and the first important moment of the game. White didn’t actually play very ambitously in the opening, but now he is presented the possibility to sharpen things. The big question therefore is: to take or not to take. Should White give up the pair of Bishops for a small structural advantage? During the game I couldn’t remember any concrete endgame-examples in which White does this. It is, of course, a well-known attacking motif to weaken Black’s king side, as in the following famous game:
Kasparov-Timman
Amsterdam 1994
18.Bxf6 gxf6 19.Rc1 Rc8? Better was 19…Bxc3. Now White starts a fierce offensive.
20.Ne4! f5 21.Ng3 Qxd5 22.a3 Bd6 23.Nxf5 Rcd8 24.Re5! and soon 1-0
The following game by kasparov is also well-known, even though he was less successful here. Again Kasparov tries to attack by mutilating Black’s structure, but Karpov defends well and won the game in the end:
Kasparov-Karpov
Sevilla (m/2) 1987
18.Bxf6 gxf6 19.Ne4 Kg7! 20.dxc4 Rad8 21.Rb3?! Nd4! -/+
In both cases White’s purpose was the direct assault.
In my game it was different. In the endgame, the bishop pair is usually equally strong as in the middle game, and also it was not clear to me how I could attack the weakened double f-pawn in the near future, let alone conquer it. Luckily I realized pretty soon that in fact I didn’t have much of a choice, since White has to give one of his bishops after 14.Bd3?! Nd5! anyway. So I quickly played (see diagram Moll-Hoffman above):
14.Bxf6 gxf6 15.Nd4 After that I managed to exchange Black’s white squared bishop, in return for an isolated d-pawn, which resulted in another difficult situation. More about that in part two. After the game there developed a lively discussion in the analysis room about this type of positions. I defended, partly to play my part, the point of view that even if I could have played 14.Bd3, I would have taken on f6, because it was the only way to play for a win. Someone else put it to me that I was taking a risk by that because of Black’s bishops. We didn’t reach a conclusion. I still wonder what rule you have to follow in these situations and I would like to ask the Grandmaster editors Karsten Müller and Erwin L’Ami if they have any thoughts on this. The structural advantage is obviously there, but there are so many positions where the bishops are even an advantage with a pawn down…
In practice, the dilemma rises frequently. There is even a line of the Exchange-Queen’s Gambit where the problem arises after only nine moves:
1.d4 d5 2.c4 e6 3.Nc3 d5 4.cxd5 exd5 5.Bg5 c6 6.e3 Bf5 7.Qf3 Be6 8.Bxf6 Qxf6 9.Qxf6 gxf6
The variation is not very popular, by the way, but the fact that strong Grandmasters want to play this with both White and Black, indicates that it’s not that clear to evaluate. Fine postion playes like Karpov, Ivanchuk and Gustafsson have played it with White, while defenders like Vaganian and Andersson have defended the Black side. It’s an interesting position to play out.
Eleven years ago, I was a live witness of the following game:
Topalov-Shirov
Wijk aan Zee 1996
After a very exciting opening (Arkhangelsk-Ruy Lopez) you’d think White would still have some endgame advantage because of the pair of Bishops. Of course you’ve already guessed White’s next move:
21.Bxf6 gxf6 22.Ra7 with some practical problems for Black. I thought the game could end in a draw any moment, but to my amazement the followed a heavy struggle and Shirov could only make a draw after move 53. A few months later, during the Eijgenbrood-tournament in Amsterdam, I got the following position on the board:
Bettman-Moll
Amsterdam 1996
After his last move (25.Rd4) White offered a draw. Suddenly I remembered Topalov’s attempts and I decided that the structure justified playing on. With 25…f5!? I tried to activate my king. After a lot of technical adventures I managed to win the game. It’s possible that ever since that game I feel that the structural advantage is tangible, even when the opponent has more activity.
But as always, there comes a moment when you start to doubt your certainties. The next game was a sobering lesson:
Rogozenko-Morozevich
Istanbul 2000
21…Qf6!! 22.Qxf6 gxf6 When we are recovered from our amazement, we can actually see the idea behind Black’s astonishing queen’s trade: White has problems developing, and Black is very active.
23.h5 h6 24.Rh4 c5 25.Be2?! (better 25.Bg4!) 25…Bb3! 26.Rf4 Kg7 27.g4 Rd6 28.Re4 Kf8 29.Rf4 a5 30.Re4 Rd8 31.Rf4 Ke7 32.Re4+ Kd6 33.Bd1 Be6!
And here White forgot to block Black’s doubled pawn with 34.Rf4. Instead, in time trouble he played
34.Be2? and after 34…f5! 35.gxf5 Bxf5 36.Rf4 Ke5 Black was so active that White soon collapsed under the pressure and lost. An impressive sample of will-power by Morozevich!
Here’s another surprising moment from again a game Kasparov-Karpov, and again Karpov is successful with his isolated f-pawn in the end.
Kasparov-Karpov
New York (rapid) 2002
White is better, but Black has a target on d4. That’s why White decides to eliminate the bishop on f6:
24.Lxf6 gxf6! What’s this? Well, after 24…Qxf6 25.Qa4! White would be fine because of the unprotected pawns on c7 and a7. Now Black is hanging on and after 25.Qa4 c6 26.Bf1 Kf8 his position was still defendable. After some mistakes by White, Karpov even won the game.
But don’t let the impression convince you that it’s always such fun to have this isolated f-pawn. Gelfand has played two instructive games that show how annoying it can be to have no pawn breaks on the king’s side.
Barcot-Gelfand
Leon 2001
16…Bxf3! 17.gxf3 Ke7 and Gelfand went on to torture his opponent for 60 moves. His knights effortlessly jumped to the beautiful squares d5 and f4 and White’s bishops where staring into blank space. Only in the far endgame he missed the win which enabled Bacrot to draw the game.
Gelfand-Grischuk
Rusland 2004
In the Tarrasch Defence, Black often has an double c-pawn, and usually it’s not such a big deal. It is here, especially after:
16.Bxf6! gxf6 17.e3 and Black not only have a double c-pawn but also the infamous doube f-pawn. Gelfand next indicated with his fine technique that Black should revise this particular opening line.
What can we conclude from these examples? It appears that activity, as ever so often, is more important  than structure. The pair of bishops can guarantee this activity, but it’s not always the case, as the examples indicate. What I think all examples do have in common is the activity of the rooks. The advantage of an isolated doubled pawn is that you have two extra open lines. But if you can’t operate on these lines, ‘what’s the use? It’s funny that rook-activity also plays an import role in the following of my game against Hoffman with which we started. More about that in the next article. Meanwhile, I’m looking forward to your ideas or suggestions in the comments!
Loading ...
[…] Read more… KAURAVAS WINNING AGAIN - January 3rd, 2007Fritz 10 Software - January 3rd, 2007Cheating allegations a blight on Indian Chess - January 3rd, 2007Full version of Oct Rating list - January 3rd, 2007Malaysia’s First Certified FIDE Trainer - January 3rd, 2007 Sponsored Links Related Reading […]
Among many things preventing me to become better player is my ungrounded unpleasantness while playing with doubled pawns and fixed pawn chains, and consequently my aim to avoid such situations without really trying to look deeper.
I can’t say anything about your game as some diagrams that have been here this morning are missing.
Sorry, diagrams are there after the page refresh.
You said: “I defended, partly to play my part, the point of view that even if I could have played 14.Bd3, I would have taken on f6″.
In the given position I would play Bd3 if I could, because I fancy (again ungrounded) pair of Bishops. Nd5-Nb4 pretty much limites white’s options.
Doesn’t the bishop normally go to g6 in that QGD Exchange Variation?
Bc8-f5-e6 is a loss of tempo and the bishop’s a bit duff on e6 imo. It could be vulnerable to a further bluntenning too from f4-f5 - since white will have Bd3 to support that push, unlike in the Bg6 line.
Surely black plays ..Bf5 not to court the bishop pair for the doubled pawns, but because otherwise white will get a perfect set up with Bd3?
Nice article,
it stimulated me to do some database search, looking at situations where one side has a bishop for a knight for at least 3 moves in a row and a double f- or c-pawn. Due to Goran’s remark I distinguished between players with an Elo>2600 and 22002600:
Black: 60-40
White: 50-50
Bishop pair and double f-pawn, both players 22002600:
Black: 78-22
White: 63-37
Bishop pair and double c-pawn, both players 2200
ok, something went wrong with the results, but here is a conclusion
What does this mean?
First of all 2600+ seem to do very well with white playing agains or handling an isolated double c-pawn while holding the bishop. Especially when we compare this with the winning percentage against or holding an isolated f-pawn. We can conclude that the white player is ussually right when he lands up with or against the double c-pawn.
The 2200-2600 category seems to have more problems as white to evaluate the case of the double c-pawn for both sides, resulting in a more balanced score sheet. In case of the double f-pawn, there is a small difference in results between both levels. Interesting is that the creation of a isolated double f-pawn, in exchange for the bishop decreases the winning chances of the double pawn side.
Therefore B*f6!, even if Bd3 is possible
Flip
Flip, I’m not sure but doesn’t this just show that 2600+ are .. just good players, no matter what position is on the board? Also, I don’t see the relevance of the results of White and Black unless you mean White is the player with Bishops and Black the position with the structure or something.
To get a good comparison, you would have to take only opponents of equal strength. Or is this actually what you did?
My suggestion would be for example to take only 2600+ players in a sample, and see how well the bishops+lousy structure are doing against good structure without bishops.
And then we can compare these results for 2200-2600 players only. Is there a difference in results? Do 2200-2600 players score better with bishops (or structure)? (My guess is that the results won’t be significant, by the way, because it depends too much on the details of the position - but then again, maybe there really is a sort of rule for this.)
It could turn out that it’s statistically better for a 2600+ player to take on f6, but for a 2200 player it’s statistically better to play Bd3
@Tom. You’re right, Black can also move his bishop to g6. I can see two reasons why Be6 also has its merits. After White can play f2-f3 and finally e3-e4 (usually preceded by Nf3-h4, g3, Bg2 and Kf2), d5 badly needs extra protection since after d5xe4 f3xe4 White has a great centre. Also, the move f6-f5 which Black sometimes has to play, is even less attractive with a bishop on g6.
So, both moves have advantages and disadvantages.
Yeah sorry, i wrote my lines in word and something went wrong in the copy-paste.
I actually did what you suggest. What I observed is that 2600+ players increase there winning probability with white is both cases (compared to average winning percentages). So when the white player in a 2600+ match accepts a lousy structure for a bishop, he increases his winning odds in most cases, while when he gives the bishop to destroy the opponents structure he also usually increases his winning odds. This might has to do with certain opening structures, for example in certain variations in the sicilian where the c1 bishop controls the important square d6. For black I cannot think of popular openings where you accept isolated c-pawns for an active bishop.
I compared these results with the 2200-2600 range. There the increase in winning odds is much smaller. This indicates a higher magnitude of random versus lousy structure/ bishop exchange, while for the 2600+ player it seems to serve a more positional goal.
However, also for the “white” 2200 players the winning percentage against the double isolated f-pawn versus a bishop was higher than average, so B*f6 is statistically a good decision.
ps. Rapanui, happy new year ey!
Happy new year to you too Flip. I still don’t see what the color-comparison tells us except that perhaps there are more openings for White where the bishops are active. It tells us, I think, more about opening theory than about chess strategy. Not to mix things too much, I suggest dropping the color-variable for now. Let’s only focus on the comparison between structure vs. bishops.
In games between 2600+ players, do you find better results (independent of color) for the structure or for the bishops?
I had a quick look at some stats on this line rapanui - on the Shredder on-line database Be6 is played 10 times, Bg6 over 100. However - Be6 scores higher - 40% compared to 36%. Although maybe 10 games is too low to draw definite conclusions.
I have played the Bg6 line quite a few times myself. I’ve not had any particular problems with it OTB or at blitz, but in correspondence games I’ve found the position almost impossible to defend. White just slowly (very slowly) builds up, black never has any sniff of a winning shot.
Anyhow.
Do you remember the game Salov-Ljubojevic from that 1994 Sicilian Thematic tournament? The two bishops were out of the question in that endgame - but black had doubled f-pawns. Salov blockaded them and then made use of his q-side majority. Maybe in your game Bxf6 can be followed by a similar piece blockade on f5, with f3, g4, h3 thrown in to follow. Although possibly black would need one more weakness somewhere else for you to exploit the blockade. Also there is a risk of BxNf5 and OCB.
Just some thoughts from a patzer
@Tom. I had forgotten about this game by Salov, but it’s very instructive. By the way, it seems to me a typical example of a position where having isolated doubled pawns actually is an advantage. Just imagine if Black could have used square e5 for one of his pieces, or the e-file for his rooks to attack the pawn on e4. But now Black was badly blocked for the whole game. It made a huge difference.
In my game I actually managed to occupy the key-square f5 as well, although at the cost of an isolated d-pawn, resulting in yet another interesting but very difficult structure-situation which I’ll talk about in part 2 of the article.
I look forward to part 2 . . . and yes you’re right - the e5 pawn is badly in the way for black in Salov-Ljubo. I’d not thought of that.
Hi there rapanui
I think the blockade strategy (Bxf6 followed by trying to play for Bf5 - to exchange off blacks light squared bishop, then Nxf5 hopefully) is the best way to approach the position - but I have not checked this with for example engine analysis. The plan has been disputed on the effectiveness of the f5 blockade at:-
http://www.chessworld.net/chessclubs/forums_thread_show_one_posteronleftstyle.asp?forumid=27&ThreadID=5930357#message5934941
However, if it can be reached, then I believe, White could then try and actually undouble black’s pawns with a later g5 which would result in two black isolated pawns in the ending on f7 and h7. It would help in my view also to exchange off a pair of rooks if possible. And most certainly try and get in bf5 to make sure black has not got the bishop-pair.
Best wishes
Tryfon (kingscrusher on chessworld.net)
Indeed, Tryfon, if White could get an untouchable knight on f5 he would be close to winning as your scenarios suggest nicely. It turns out, however, that this is not really possible because after Bf5, Black always plays Bc5! challenging the knight on d4 (and the pawn on f2).
In the game I did manage to exchange light-squared bishops by means of Bf5xd7, but at the cost of an isolated d-pawn (Bc5xd4). This led to another very interesting structure about which more in part 2.
Why should the knight reach f5 via d4 though, and not h4? Nh4 also has the merit of defending pg2 against Rg8, until a supported g4 is possible . . .
@Excellent suggestion, Tom. I completely missed this idea, since I wanted to play for the trick Nxc6 (which by the way is no trick at all.) However, things are still not so simple. If White plays 15.Nh4 instead of 15.Nd4, then after 15…Be6 Black can probably still challenge the blockade on time with Bh6-g5. However, things still look OK for white there, so perhaps I should have played Nh4 anyway.
Hoi Arne,
Na zo’n oproep kan een reactie natuurlijk niet uit blijven
Interessante stelling tegen Hoffman. Na 14.Lxf6 gxf6 15.Pd4 lijkt het me objectief remise. Allereerst vroeg ik me af hoe het staat nu 15…Le6 16.Pxc6!? f5 17.Lf3 Lg7 18.O-O-O bxc6 19.Lxc6+ Ke7 20.Lxa8 Txa8 21.f4. Moet ok zijn voor zwart: a5-a4 en Tb8 geeft zwart tegenspel.
Anders kan zwart ook 15…Lc5!? kiezen (16.Pxc6 O-O!) bijvoorbeeld 16.O-O-O Le6 17.Pf5 O-O gevolgd door Tfd8 ziet er ook weer ok uit voor zwart. Kortom, ik denk dat de dubbele f-pion te weinig is om echt van voordeel te kunnen spreken en dat zwart weinig gevaar loopt.
Hi rapanui
There are some ideas for White which boost the blockade strategy idea to make it a lot more playable including castling queenside, and also being willing to sac the pawn on f2, if black plays later Bc5.
Here is some concrete analysis to check the blockade plan, with the help of Rybka on very quick time given to it:
1. Bxf6 gxf6 2. O-O-O Bd7 3. Rhe1 O-O-O 4. Nd4 Kc7 (4… Bd6 5. Bf5 Bf4+ 6. Kb1
Be5 7. Bxd7+ Rxd7 8. Nf5 Rxd1+ 9. Rxd1 Re8 10. g3 Bb8 11. f4 a6 12. Rd2 Bc7 13.
c3 Re1+ 14. Kc2 h5 15. h4 Re8 16. Kd3 b5 17. Re2 Rh8 18. Re7 Rd8+ 19. Kc2) 5.
g3 Bd6 6. Bf5 Bxf5 (6… Bc5 7. b4 Bxd4 (7… Bxf5 8. Nxf5) 8. Re7) 7. Nxf5 Bc5
8. b4 Rxd1+ (8… Bxf2 9. Re7+ Kb8 10. Rxf7) 9. Rxd1 Bxf2 10. Rd6 Be3+ 11. Kb2
Rf8 12. Kb3 Bg1 13. h4 Bf2 14. Rxf6 *
I find clear advantages for White in all variations.
I prefer Nd4 to Nh4 because it offers greater flexibility, and didn’t really see the issue you have based on Bc5, but maybe I will check the analysis later again. As Tom saids, Nh4 is also possible, and this was my original “implementation” of the f5 blockade strategy.
In the conclusion of this article you talk about structure vs activity. I would also consider the actual ability to blockade being function of the effectiveness of doubling the opponents pawns. As Nimzovich has told us “Restrain..Blockade…then destroy!”. So if there is only the structural damage part, without these three parts, then the doubled pawns or structural damage may be more academic in nature, and not so “exploitable”.
The above analysis is illustrated at:-
http://tinyurl.com/yhe5fw
Best wishes
Tryfon Gavriel
Fide 2157
Hi Tryfon,
I think you’re too optimistic about the position for White. For one, I think the opposite-bishop ending you mention is quite an easy draw for Black, and even if White can get a perfect blockade with his N on f5, it’s still not clear. The real problem is that White can probably never really get this ‘ideal’ position due to concrete moves.
For example, after 1.Bxf6 gxf6 you suggest the normal-looking 2.0-0-0 but now I was afraid of simply 2…f5!? with the idea Be6 and Bg7 and I think White may actually be worse here.
I agree with Erwin that it’s probably an equal position after Bxf6. Black simply has too many options to disrupt White’s simple plans. Still, White may be worse if he doesn’t take on f6, so that justifies it entirely.
I’m not quite sure I understand that?
Then again it’s probably just me.
I like your site. Well done!
I’ll come by again soon.
Very nice.
Keep up the great work!