Sunday, November 4, 2018

Game Analysis: Balancing Attack and Defense

Do you ever wonder why many GMs always seem to find these eye-popping attacks that work, but you try to do the same thing and it blows up in your face? Was there this miracle defense that you overlooked? Was there one move that was not check that actually allowed your opponent to mate your King in a long serious of continuous checks? The next time that you think about "going for it", did you shy away from it because of your past experiences, just to find out afterwards via artificial intelligence that this one would have actually worked? Did that shying away turn a win into a draw, or even worse, a lengthy downhill plunge into a loss?

The game that we are going to look at today shows a combination of positional ideas and tactical combinations where White will gain a winning advantage after failure by Black to properly judge whether to sit back and defend or try to counter-attack White, turning a "+1" position into a "+4" position for White. This "winning position" will not be followed up with the "best move" every time, and quite frankly, there is no reason to feel like one must find the best move every time. We will see White being a little overly cautious, but not to the extent that he just gives up on the attack and allows Black to come back and draw the game. The thing to keep in mind is that the main goal is to win the game, and the key is never to give Black the opportunity to come back and equalize the position, but whether the position is +7 or +3, and whether you win the game in 8 more moves or 14 more moves is 100% irrelevant. The way to judge whether you conducted a successful attack is by determining, once you have achieved a winning position, did you ever allow the game to return to equal or "a slight advantage for White"? If not, you conducted a successful attack. If so, your attack failed, regardless of whether you ultimately won the game or not.

Let's take a look at the game where we are going to see White reach a winning position, and once it's reached, it is never given up, despite not playing the best computer move every time.

NOTE: The player playing Black is a different "Andrew Jiang", from Georgia, and not the local player form North Carolina


Washington Chess Congress, Round 4
W: Patrick McCartney (2069)
B: Andrew Jiang (1765)
Owen's Defense

1.Nf3 b6 2.e4 Bb7 3.Nc3 d6

While this is not an opening article, I will point out here that this move is not good. It severely weakens the light squares. If you think about various hypermodern defenses where Black fianchettos a Bishop rather than fighting for the center via central pawn advancement, you will notice that the first two pawns advanced are typically three files apart. This is not a coincidence! If you take the starting position of the eight pawns, and observe any two that are three squares apart, you will notice that they reside on opposite colored squares, releasing opposite colored diagonals for the two Bishops. The Knights and the Queen have multiple ways out, and the Rooks are the last pieces developed in most cases, and so the initial advancement of the pawns requires diagonals of both colors to be opened in order to properly develop all of your pieces. This is why you see the d-pawn and g-pawn advanced in the Pirc, King's Indian, Grunfeld, and Benoni lines while you see the e-pawn and b-pawn advanced early in the Nimzo-Indian, Queen's Indian, English Defense, and Owen's Defense.

The move played weakens the light squares and overloads the dark squares with pawns and does nothing to help the Dark-Squared Bishop get out.

Instead, Black should play 3...e6, intending something like ...Nf6 along with ...Be7 or ...Bb4, a lot depending on how White responds. For example, if White plays d4, the Bishop may be best placed on b4 while if White develops his Kingside first and castles, holding off on d4, then the Bishop is probably best placed on e7.

4.d4 Nf6 5.Bc4


Did White just drop a pawn?


5...e6

Already we see one problem with having played 3...d6 instead of 3...e6. Had Black played 3...e6, White's 5th move would not be possible and he'd have to defend the e4-pawn. However, here, taking the e-pawn is a tactical blunder. After 5...Nxe4?? 6.Nxe4 Bxe4 7.Bxf7+! Kxf7 8.Ng5+ and 9.Nxe4 and already White has a winning position. No material was gained, but Black is positionally busted already here!

After 5...e6, we see another problem with 3...d6. The Bishop on g7 is severely limited. It can no longer go to b4, and has to resort to the passive e7-square. This gives White a free hand at the center.

6.Qe2 Be7 7.d5 e5?!



Black should trade on d5 and then immediately castle to get out of the pin. After 7...exd5 8.exd5 O-O, White has a slight advantage due to the space advantage, but Black eliminates a White central pawn.

The other aspect of the position is a positional idea that favors White due to a tactical solution. We see that both sides have a bad Bishop. White's Light-Squared Bishop and Black's Dark-Squared Bishop. While it is going to be difficult for Black to solve his problem, which will be seen by the fact that the final position still features the bad Bishop for Black, the fact that White does not have to respond to Black's move since it doesn't involve a need to recapture, and since the Black King still sits on e8, White has a tactical way to eliminate his worst piece, and how Black responds will lead to another positional weakness for Black.

8.Bb5+ c6

This will lead to a major positional weakness for Black. Do you see what that weakness will be? Other moves don't help. If 8...Nbd7 or 8...Nfd7, White should not be quick to trade off the bad Bishop for the Knight. Instead, 9.a4 and if Black plays 9...a6 (here or later on), White should retreat the Bishop to d3. Yes, he keeps his "bad Bishop", but Black has weakened his queenside structure. He has to watch out for a6, keeping the Bishop and Rook immobile. b6 is weakened, and if he ever tries to undermine with ...c6, then the b6-square will be very weak.

9.dxc6 Bxc6 10.Bg5!

When trying to figure out Black's weakness after 8...c6, did you consider this move? White did not want to trade his Light-Squared Bishop for a Knight because the Black counterpart would be too strong, but when you consider the opponent's bad Bishop, in this case the Dark-Squared Bishop, it is often a good idea to trade your own "good" bishop for a Knight in closed positions as it leads often to the "Good Knight" (hint hint about Black's weakness achieved from his 8th move) versus Bad Bishop scenario.

10...O-O 11.Bxc6 Nxc6 12.Bxf6! Bxf6 13.Nd5

And now we see what White was after following Black's 8th move. This is the classic pipe-dream scenario in closed positions. Open up an outpost, which White did on move 9, and then trade away both of your Bishops. The Bishop that is the same color as the outpost needs to be traded for the opposing Bishop, and the Bishop on the color opposite that of the outpost needs to be traded for a Knight, leaving White a "piece up" from the aspect of being about to control the outpost, in this case, d5. Both of White's Knights can get to d5 at some point. The Black Bishop will never be able to contest the d5-square because it's the wrong color Bishop. In addition, because of the Black pawn on d6, it's a lot harder for Black to counter the d5-square with his heavy pieces. White can use the semi-open d-file. White has a significant advantage here.

13...g6 14.c3

And here is the first example of where a move does not need to match the computer's "top choice" to be classified as a good move. Does the move lower the general assessment? After Black's 13th move, the position is "+/-" (Clear advantage for White). After White's 14th move, it's still "+/-". The fact that it's a small fraction lower does not make the move bad.

White's idea is simple. It has to do with timing. Black has no good way to contest the d5-square anyway, and so White takes two key squares away from the c6-Knight, namely b4, and more importantly, d4, since it would plug up the file and stop White from using the heavy pieces to continue to control d5 and attack d6. If Black tries to play 14...Ne7, the Bishop hangs, and if it didn't hang, a move like 15.Rd1 or 15.O-O-O continues to dominate d5. The one thing White must make sure about is that Black is never able to trade on d5 in such a way that White would be forced to re-capture with the e-pawn. White can exit the outpost as long as Black can't respond with ...d5 successfully, maintaining the backwards pawn and the "cavity" on d5. If Black can trade on d5 in a way that forces White to take with the pawn, that White pawn on d5, which shields the weak d6-pawn, would be the equivalent to being a "filling" for the cavity on d5. We are here to brutally attack the Black King, not to serve as his dentist!

14...Bg7 15.h4

And now White proceeds to play the move that artificial intelligence wanted White to play the previous move. The idea is simple. The move 13...g6 created a hook for White. The fact that White has not castled serves in his best interest as the Rook on h1 will be very strong if Black allows 16.h5, and so his next move is virtually forced.

15...h5 16.Rd1

There is always the question here, 16.Rd1 or 16.O-O-O? It turns out both moves are fine. But White has to make that ultimate decision now as there is no other waiting move to get more information. The Rook needs to go to d1, and so White must decide now whether he wants the King on the Queenside or in the Center. The odds of White ever castling Kingside is very slim in this position.

16...Qd7



So here we have a critical moment. White wants to attack the weak d6 pawn. White also has to watch out for desperation counter-attacks by Black with moves like ...Qg4. In order to get at d6, White will eventually have to move the Knight, presumably to e3. White will also need to relocate his Queen as it is misplaced on e2 at the moment as there is no pressure on e4 like there was in the opening phase of the game. Which should White do first?

17.Qb5

White brings the Queen out first! To figure out whether or not this is successful versus 17.Ne3, a few things must be factored in.

First, what does 17.Qb5 do for White? Well, it continues, for now, to prevent Black from contesting d5 with the Knight as the Knight is pinning to the Queen, and so 17...Ne7 is impossible. It also brings the Queen closer to a square that is closer to the weak d6 pawn, where ultimately, the Queen will likely go to d5, or in some cases, b4 if the Knight eventually moves. Note that 17.Qd3 would immediately hit the d-pawn once the Knight moves, but it does not tie down the Black Knight like 17.Qb5 does. Also, it is hard to harass the Queen as advancing the a-pawn weakens the Queenside pawns.

On the flip side, what does moving the Queen away cost White? Does Black have anything that causes a major problem? The only move that could possibly make 17.Qb5 bad is 17...Qg4, hitting both e4 and g2. Let's say White grabs the loose Knight with 18.Qxc6. Does Black have a counter that makes this bad? Well, we can immediately rule out 18...Qxg2 as the move 19.Ke2 protects both the Knight and the Rook (via the Rook on d1). That leaves 18...Qxe4+. Can White get out of this and maintain a winning position. It turns out there are two ways for White to do it:
  1. 19.Kf1 Qc2 20.Ne7+ Kh8 21.Qxd6 Rae8 (21...Rad8 loses to 22.Qxd8 Rxd8 23.Rxd8+ and now 23...Kh7 24.Ng5+ Kh6 25.Ng8 is mate while 23...Bf8 24.Rxf8 Kg7 25.Rg8+ Kf6 26.Nd5+ Ke6 27.Rd8 and White can answer any back rank check with Ne1, giving White two Rooks and two Knights for the Queen) 22.Kg1 and now 22...Rxe7 fails to 23.Rd2 Qb1+ 24.Kh2 Qf6 25.Qxe7 while 22...Qxb2 23.Ng5 Qxa2 24.Rh3 Qc2 25.Re1 sees Black getting only two pawns for the piece.
  2. 19.Kd2 also works. After 19...Bh6+ 20.Ng5 Bxg5+ 21.hxg5 Qxg2 22.Rhg1 followed by 23.Qxd6 and White is winning.

Therefore, there is no need to allow 17...Ne7 by playing 17.Ne3 right away, and 17.Qb5 is therefore the better move here.

17...Kh8 18.Ne3 Rac8 19.Qd5

Once again, White doesn't play the top choices by artificial intelligence, but the advantage still remains for White because this move also has a specific point. Of course, the immediate 19...Ne7 drops the d-pawn, and after the game move, the f-pawn pawn is weakened, and so once again, just because it's not the top move doesn't make it a bad move as long as there is a specific point behind the move played, and we will see that White attacks both weaknesses for Black, namely d6 and f7.

19...Rfd8

Adding defense to one weakness (d6), but further weakening the other (f7).

20.Ng5 Rc7 21.Nc4



So here is a defensive test for Black. What is Black's best move here? Should he sit back and defend? Or should he try to counter-attack White's weaknesses?

21...Qg4?

This is the wrong answer. Black had to acknowledge that he is losing a pawn as the pressure is too much, and solidify his position. Theoretically, Black will still be lost, but the only way to force White to prove himself and give White any difficulty in winning the game, he needed to play 21...Bf8 22.Rh3 Qe8 23.Rf3 Rdd7 24.Rf6 Kg7 25.Nxd6 Bxd6 26.Rxd6 Rxd6 27.Qxd6 Re7 where the assessment teeters between White having a significant advantage and White winning, but White still has a lot of work to do as all he has here is an extra pawn in the form of a queenside majority and neither King is exactly safe.

The move played in the game leads to disaster.

22.Nxd6 Qxg2 23.Rf1

And here we have another example where White exercises more caution than necessary, but in no way does he allow Black to come back. As long as he doesn't overdo the defense to the point that Black is able to come back, there is no real issue in not finding the computer move that leads to all the fireworks. 23.Ngxf7+ Rxf7 24.Nxf7+ Kh7 25.Nxd8 Ne7 26.Qd3 Qxh1+ 27.Ke2 Qxh4 28.Qf3 is the line the computer gives with a +5 position for White whereas the move played only leads to a +1.9 position after 23...Rf8 24.Qc4, but in either case, White is clearly winning, and so this is not the time to think that you screwed up simply because your moves don't match that of Carlsen or artificial intelligence. The important thing is that Black is still not surviving! Of course, Black doesn't play 23...Rf8 here, and so his position will end up just as bad as it was in the computer line.

23...Kg8 24.Ngxf7 Rxf7 25.Qxf7+ Kh7 26.Qc4

Once again, White doesn't play the "computer line" of 26.Qc7 Rxd6 27.Rxd6 Qxe4+ 28.Kd2 Qf4+ 29.Kd1 Qa4+ 30.b3 Qxa2 31.Qxc6 and yes, the White King squirms around a little and White ultimately wins, but again, humans are not computers, and if you find a winning line, like what White played here, why kick yourself for not finding the computer move? Again, as long as your move continues to be winning, and not "slightly better" or "equal", there is nothing ultimately wrong with the move played.

The rest of the game is a matter of technique, and Black puts up little to no resistence.

26...Na5 27.Qe2 Bf8 28.Nf7 Rxd1+ 29.Kxd1 Bg7 30.Ng5+ Kg8 31.Kc2 Bf6 32.Nf3 Qg4 33.Rg1 Qe6 1-0

And here, Black resigned during White's 34th move. The game would likely continue 34.Rxg6+ Kf8 (34...Kf7 or 34...Kh7 drops the Queen to 35.Ng5+ while 34...Kh8 allows mate in 7 starting with 35.Qd2 or 35.Qe3) 35.Ng5 and now 35...Qc6 allows 36.Qxh5 while 35...Qxa2 allows 36.Rxf6+ Ke7 37.Rf7+ and Black must either give up his Queen or else go to a square that allows White to check with the Queen and once the Queen gets in the game, it's over for Black.


So what we saw here was a well-conducted attack by White that maybe didn't match the level of magic that Rubinstein showed in 1907 against Rotlewi or Carlsen showed in the final playoff game against Karjakin with the Queen Sacrifice, but White was able to maintain the more sensible opening play, kept the big advantage in the early middle game, and maintained a winning position throughout the second half of the game. If you are able to do this, you should consider your game a success. In the digital era, there are way too many players, especially younger players, that rely way too much on artificial intelligence. Many of them try to use artificial intelligence as their only form of studying, which typically fails, and judges most of their moves solely on whether artificial intelligence agrees or not, and if it's not the top choice given by the machine, they often think they failed, and that is not the case.

The game shown above should not be confused with games where the assessment constantly bounces back and forth from +/= to -/+ to +- to =/+ to +/- to a win for White. You win a game like that and yes, you did indeed win a game that was not played well at all, but what is often missed is this fine line in between. Not all games played fall in the category of "computer perfection" or "sloppy". There is a very wide area in between, and a large portion of that area can still be deemed a "successful attack". This game featured a successful attack by White and failed defense by Black, particularly at move 21 after the bad opening play. The difference between a GM and an Amateur is not GMs find the fireworks and amateurs don't. The games with fireworks are the ones that hit the books. For every one of those, the GMs have played and won numerous games without fireworks. The real difference is that GMs are able to execute successful attacks like this one on a consistent basis whereas an amateur will often follow up the strong play by White in this game with a complete dud, whether it be the very next game he plays, or some upcoming game in the near future.

The moral of the story is that working on your game and looking for consistency is far more important than wasting your time looking for glorified fireworks. If they pop up, great, but they are not the only good moves that can be played on the chess board.

Until next time, good luck in your tournament games.

No comments:

Post a Comment