Sunday, June 10, 2007

"Chess Narratives" Exercise

Black to play after 15.Qb3.
What's Black's best move and why?

In Chess for Zebras (reviewed here last year), Jonathan Rowson talks about how chessplayers try to make sense of the middlegame with what he calls "chess narratives":

I think of chess narratives as the background 'noise' that permeate our thoughts during play and this 'noise' is often sufficiently loud that it operates as the context of our thoughts. For instance, if you probe the advice 'counter an attack on the wing with play in the centre' for a few seconds, you can imagine someone telling a story about the game, with that as the basis of the plot. Narratives are the guiding stories that give us a sense of what we are trying to do and why. ... While nobody is immune to these narratives, very few use them to guide themselves towards correct moves. Many players get lost in these stories, trapped by their own narrative, and completely lose track of the objective state of affairs on the board... (46).
The best narrative offers a good reading of the position that fits well with specific calculations so that, as Rowson puts it, "your assessment and your variations make sense of each other." A bad narrative, though, is really a self-deceptive invention (what Rowson calls "fabulation") that is imposed on the facts with little close inspection of the specific lines that might support it. Like "The Lazy Detective," the fabulist cares less about the truth than in confirming what he already knows based on his ideological biases. More often than not, however, the truth will come out and the fabulist will fall--though a skilled fabulist will then be able to explain away his failure without acknowledging the truth (rather like George Bush discussing Iraq).
In my game with John Moldovan (The Chess Coroner) from the KCC Summer Tournament (which began this week--though you can join at any time), I struggled to find the right narrative at a critical juncture. I think I made a more accurate choice than I usually do, but not necessarily the best. Let's see if you can do better....
Which of the following chess narratives best fits the position diagrammed above and leads to the best choice of plan?
a) 15...Qh5
Narrative: Black has a tremendous lead in development, of at least three or four tempi, which justifies sacrificing at least a pawn to gain an attack. White's King is also practically denuded of defenders, while Black's forces are ready to leap to the kingside, supported by the pawn at e4. Even his Rook is ready to swing over to support the Queen and Knight. Black can target both the h2 and g2 squares, but the most promising idea is to use the Queen and Rook together in a full frontal battering-ram assault on g2. The basic idea begins 15...Qh5! 16.Qxb2 (if 16.h3 then 16...Rg5! 17.Kh1 Rg6! followed by ...Qg5 to batter down the door) 16...Ng4! 17.h3 Rg5!! and all of Black's forces are launching themselves at White's helpless King. Not only that, but Black gets to finish things off beautifully after the natural 18.hxg4 Qxg4! 19.g3 Rh5!! and there is no defense.
b) 15...Qd6
Narrative: Yes, Black's lead in development dictates that he attack the enemy King, but the best target of attack is h2. The idea is 15...Qd6 16.Ba3 (who can resist pinning the Rook to the Queen?) 16...Ng4! 17.g3 Qh6! 18.h4 and now Black can attack by 18...Rf5! followed by ...Ne5-f3+ and ...Nxh4 to blast open the weakened White King position. Meanwhile, White's forces stand helplessly by on the queenside. He's dead meat.
c) 15...Qxb3
Narrative: When I play chess, I always ask myself, "What would Capablanca do?" Sure, there are chances of attack on the kingside, but that all seems very speculative to me and seems to risk losing Black's real advantage, which consists of his control of the c-file, his queenside pawn majority, and his better minor piece. The best way to capitalize on these positional pluses is not to throw away material (by allowing Qxb2) seeking an attack, but to simplify into a winning endgame by exchanging Queens. After 15...Qxb3 16.axb3 a6 I will double Rooks on the c-file, play my Knight to d5, pawn to f5, and march my King to the center. He has almost no counterplay and will likely be forced to exchange Rooks on the c-file, leaving me with a winning endgame due to my superior minor piece and queenside pawn majority.
d) 15...Qc6
Narrative: Black's advantage consists of control of the c-file, plain and simple. I intend to triple my heavy pieces on that file and lord it over my opponent with 15...Qc6 16.Bb2 Rec8.
e) 15...Qd7
Narrative: Black needs to avoid the exchange of Queens lest he risk losing too much attacking force to exploit his lead in development. After all, a lot of pieces have been exchanged, and if Black is going to organize an attack he will need attackers. Meanwhile, Black also needs to restrain White's backward d-pawn, which he can do by 15...Qd7. This is the most versatile Queen move: it avoids the exchange of Queens, stays on the d-file to restrain the weakling at d2, defends the b-pawn, and even eyes White's kingside. I will follow up with ...Rd5, ...b6, ...Rd3, ...Rc8, and ...h6 to create an invulnerable position while squeezing White till he coughs up a pawn.
See my notes to the game for the answer (or what I think is the right answer anyway). I include the PGN file if you want to do your own computer analysis to confirm it.

If you enjoy this little "test," then I suggest you pick up the marvelous book Test Your Positional Play by Robert Bellin and Pietro Ponzetto which basically presents 30 exercises along these lines.


Tom Chivers said...

There is an interesting implied narrative beneath each of your moves: you must do something with your queen because of the tension introduced by Qb3, so you don't consider moves that just leave her there, despite the fact she is defended twice!

Imo if there's not a concrete solution on the kingside, then white's weakness would seem to be the d2 pawn, which apparently ties the bishop to c1, this giving black free-reign over the c-file and d-file. Black's majority on the a-file and b-file also looks mobile compared to d2, white's 'extra' pawn so to speak. However I can't find a way to exploit d2 - white seems to have ways out - eg Bb2/c3/d4/xf6. Ideally I think you would just stick your knight on d3 and the game would win itself.

Michael Goeller said...

Good point, Tom. I actually did not consider other moves closely and had not thought much about why that is so. But it is sort of the central question in the diagram: should Black exchange Queens or not? If no, you gotta move the Queen. If yes, why not exchange right away? You may be right that there is no need to institute the exchange since White will likely exchange himself if allowed, helping to bring a Rook or Knight to d5. Maybe it's trying too hard to force things when Black can just do something useful: Maybe double Rooks on the c-file by ...Rec8. Maybe ...b5 to get those doggies rolling. Maybe ...b6 to free the Queen from guard duty at b7. Maybe ...Rb5 to push White to exchange on d5, which may be better than opening up the a-file for him by ...Qxb3. None of these moves is significantly different from the most forcing choice for exchanging Queens, 15...Qxb3.

Getting the Knight to d3 would be a tall order, of course (no doubt you are half in jest) -- you don't want to give White the chance for 15...Ng4 16.f3 and he starts to free himself.

There are definitely other moves than those I list--granted. But I think those listed are all reasonable and suggest the different plans available to Black -- both good and bad -- pretty well.

Goran said...

Positions like this are highly uncomfortable because player would feel he has to do something concrete or his advantage will disappear.

I would probably dismiss kingside attacks without looking too deep. Certainly not all the lines you wrote in the commentary. White has f4 and some Qb7-c7 and that seems like enough of defending force.

Qc6 d4 looks like permanent small advantage for Black, but I wouldn't be happy with that. Persistent defender would probably hold a draw. Qd7 I don't like because I would think that's natural square for N to move to c5 later.

Therefore, Queens exchange looks good enough. "15... Qxb3! 16. axb3 a6 would have been Capablanca's preference, after which Black is in control with a queenside majority and better structure. Play might go 17. Bb2! Nd5 ( maybe 17... Rc2 18. Bc3 Nd7 19. f3 Nc5 [black is slightly better] ) 18. Rfc1 Rec8 19. f3 Rxc1+ 20. Rxc1 Rxc1+ 21. Bxc1 f5! [black is slightly better] and the Knight is better than the Bishop." - This line, unless Black has stronger move in between, I don't like. Looks like White can play quick Kf2-e2 and d3. While Black Knight is still dominating, White might be in position to create kingside passer (later). Still, all this has to be analyzed over the board.

I would probably think along Rec8, Rd8, b6 or b5. Let White trade on d5. On Bb2, if there is nothing better, Black can play Nd7, which will be useful later for transferring N to d3.

In this position, I think I would be faced with "choose-among-many-solid-moves" problem, rather than with narratives about Capa or kingside attack. What is the subtle difference between Rec8 and Rd8? Pawn to b6 looks safer, but b5 might impose new problems for White? Things like that...

Michael Goeller said...

Goran --
Thanks for the great feedback. I think you hit the nail on the head when you say it's a "choose among many solid moves problem"--which is what took me the most time on the clock. That's the main pressure on the player with advantage (as Yermolinsky talks about): you have to try to figure out the best continuation, which is not easy.

The position I diagram probably works better for what Dan Heisman likes to call a "Stoyko Exercise," where you try to understand everything about the position to appreciate the rich range of possibilities that you might typically overlook or ignore. There are a lot of possibilities, as both you and Tom point out. I still think the right idea, though, is to head for the ending Capa style.

My natural tendency, however, has always been to prefer the attacking idea -- even if it looks suspect, as is the case here. So I was rather self-satisfied that, despite spending some time to explore the kingside options, I decided on something a little more sane--if not finally the best.

Goran said...

Being regular reader, I noticed your preference for nice tactics :) I have tried to simulate my thinking process in those 5-15 available mins. For time-management purpose, I wouldn't go deep into calculating direct attack (after seeing f4 and Qc7), even if the lines are very nice. Capa method demands long accurate calculation and correct evaluation, and that's a slipping field for me :) I also tend to keep open as many options as possible, which might be horrible weakness, because I could skip some good opportunities just because I don't want to "commit".

Still, I wouldn't dare to claim which move is the best. Remember Gulko-Rowson chat?

I am very curious to hear about moves-elimination from titled players, maybe you can ask FM Stoyko for his opinion.

Tom also has an interesting item on his blog