17 December 2016

Training quote of the day #10

From Peter Zhdanov's Yearbook of Chess Wisdom.
In chess, one shouldn't be too dogmatic. I had a weakness of relying too much on theoretical assessments. For example, if I realized that my position should be winning, I used to lose interest in the game and hope that it will end soon by itself. Similarly, I didn't put up the best resistance in the positions which I adjudicated as lost for myself. WGM Natalia Pogonina provided me with useful advice in this regard. According to her, you shouldn't dwell too much on the mathematical assessment of the position. Instead, it makes more sense to try to improve the life of your pieces one-by-one. If you are ahead, it will usually lead to victory. If you are behind, it might give you some counter-chances if the opponent's play is not precise enough.

11 December 2016

Commentary: 2016 Olympiad Round 7 (Shankland - Sethuraman)

This next commentary game is from round 7 of the 2016 Olympiad in Baku, which saw the United States defeat India 3.5-0.5.  GM Sam Shankland contributed to that by winning the White side of a Slav Defense against S.P. Sethuraman; original ChessBase news and analysis can be found here.  As a Slav player, I found the game interesting; I feel it's important to study losses in your openings, not just wins, which can be common to focus on.

The below game highlights a number of useful chess themes, but it's also a lesson in the value of persistent defense when under pressure, as well as how having the advantage can slip away into a loss.  Shankland was thrown on the defensive after grabbing a pawn and then running out of threats.  However, Black - under time pressure, apparently - missed several follow-up moves that would have more directly converted his advantage, for example around moves 29-30.  Although technically lost (according to the engine), Shankland kept playing effective defensive moves that helped take away Black threats, until the tide turned around move 34.  By the time move 40 was reached, it was White who had the initiative and winning threats, although the win was not assured.  The long queen and minor piece endgame is also instructive to see, both for the principles involved and for the interesting tactic 65...Bg5 which looks like it could have held for Black.

As a final introductory comment, when looking at these types of master games, it's always useful to remember the pitfalls of computer analysis and see why the top engine moves aren't made on the board, which helps improving players both better understand the game and demonstrate how practical choices often need to be made at the board, rather than always striving for an "optimal" move selection.

Shankland, Samuel L (2679) - Sethuraman, S P. (2640)

Result: 1-0
Site: Baku
Date: 2016.09.09
[...] 1.d4 d5 2.c4 c6 3.¤f3 ¤f6 4.e3 ¥g4
4...¥f5 is the other main choice here to stay within the Slav Defense. It's largely a matter of taste which to choose, although this variation offers White an easy route to exchange the Nf3 for the Bf5, if desired.
5.cxd5 cxd5 after the exchange of pawns, Black now has to worry about the e8-a4 diagonal. However, the trade-off for White is reducing the central tension and opening the c-file, which Black usually can find useful. 6.¤c3 e6 White scores a remarkable 68 percent after this move, according to the database. It's remarkable because it doesn't seem warranted with the solidity of Black's position. 7.£a4+ taking advantage of the open diagonal to harass Black. 7...¤bd7
7...¤c6?! lets White pile up the pressure and has not been played in the database. 8.¤e5 ¦c8 9.¥b5 £b6 and now 10. f3!? or 10. b3 look good for White, who can play comfortably on either the kingside or queenside.
8.¤e5 the difference here from the previous variation is that while the knight on d7 is still pinned, it is amply defended. 8...a6 a prophylactic move to take away the b5 square from White's bishop (or knight). 9.f3!? only played once before in the database (a White loss), but Komodo has it near the top of its choices.
9.¤xg4 is a more standard choice, with a knight for bishop exchange. 9...¤xg4 however, this takes the pressure off Black.
9...¥f5 10.g4 the (only) logical follow-up to White's previous move. The kingside space advantage is real, but Black need not panic. However, it requires careful assessment and calculation to select the (only) reply that keeps the balance. 10...¥g6?! this obvious move allows White to benefit from his space advantage and keep pressing.
10...b5 counterattacks immediately, to good effect: 11.£d1 shifting the queen toward the kingside action 11...¤xe5 another counterattack 12.dxe5 ¤xg4 definitely not an obvious move 13.fxg4 £h4+ 14.¢d2 ¥xg4 15.£e1 £g5 and the engine assesses that Black has full compensation in an equal position. It is certainly more fun to play the Black side here.
11.h4±11...b5 still the best idea, but one move too late to preserve Black's game. This is a common phenomenon found in analyzing my own games, as well. 12.£d1 b4 Black must keep up his counterplay on the queenside, as White is much better equipped to play on the kingside. 13.h5 Shankland correctly presses his own plan and ignores Black's threat. 13...¥xh5 with nowhere else to go, the bishop's best move is to kamikaze while the Nc3 remains under threat. (13...bxc3 14.hxg6 ¤xe5 15.dxe5 ¤d7 16.gxh7±) 14.¤xd7 this looks like the easier way for White to proceed, eliminating the idea of ...Nxe5. The engine instead suggests moving the Nc3 out of harm's way first, for example:
14.¤e2 preparing for a kingside transfer 14...¥g6 now Black seems to have gotten away with taking the h-pawn, but... 15.¤xg6 fxg6 forced, due to the pin on the Rh8 16.¤f4 ¢f7 to protect the g6 pawn, again because of the pin 17.e4± and White has more than enough compensation for the sacrificed pawn, with Black's king in an awkward position.
14...¤xd7 15.¦xh5 choosing to have the semi-open h-file and an active rook.
15.¤xd5 would mirror the bishop's kamikaze efforts; Black could then continue with the same tactical motif: 15...¥xg4 16.¤f4 g5 17.fxg4 gxf4 18.£f3 fxe3 19.¥xe3² and Black keeps the extra pawn, but White has compensation with better development and (probably) king safety.
15...bxc3 16.bxc3 £c7 targeting the backward c-pawn. 17.¥d2 although White has the two bishops, this doesn't seem to be an advantage here, as their scope is currently limited. 17...¥d6 18.¥d3 ¤b6 eyeing the c4 square. 19.¢e2 clearing the first rank for White's heavy pieces and getting off the h4-e1 diagonal. 19...h6 this turns out to be rather loosening of Black's kingside and to give White an easy target, although technically speaking it is not a bad move. Other good options: (19...¦b8) (19...g6) 20.g5 Black still has the problem of the pin on the Rh8. 20...¢d7 connecting the rooks and eliminating the pin problem. 21.gxh6 gxh6 Black's h-pawn is now passed, but also a middlegame target. 22.¦b1 ¦ag8 it's always difficult to select from different plausible-looking placements of a rook. Perhaps Black had the intent of provoking White's next move. (22...¦ab8!?) 23.¥xa6? Shankland must have not seen a way for Black's resulting attack to bear fruit here.
23.f4 is preferred by the engine, which would block the Black dark-square bishop and also better prepare the capture on a6.
23...¦g2+−⁠+24.¢d3 (24.¢f1 ¦hg8 25.¦h1) 24...¦a8 switching to offense along the a-file. 25.¥b5+ ¢d8 26.¦xh6 ¦xa2 White remains a pawn up but his king is in an awful position and Black's rooks on the second rank are strongly placed. 27.¦h8+ ¢e7 28.¦e8+ ¢f6 White is now out of threats. 29.¥e1 ¢g7 stopping ideas like Bh4, but
29...¥g3!? would get the Black bishop into the attack and remove a key White defender. This looks like the simplest way to proceed.
30.f4 blocking out the Black bishop. 30...f5
30...¤c4!? would (again) bring another piece into the attack, with strong threats.
31.£b3 Black is still winning here, but has yet to make a knockout move. White meanwhile is doing his best to contain Black's threats and generate some of his own, making winning continuations less obvious to find. The text move for example now makes the e6 pawn vulnerable. (31.¦xe6??31...£c4+ 32.¥xc4 dxc4#) 31...£f7 protecting the e-pawn, but now removing the sacrificial tactic on c4.
31...¦h2 is a subtle continuation, seizing the h-file and setting up the threat of ...Ra3 with a deflection tactic, for example: 32.¦xe6 ¦a3 and now 33.£xa3?? runs into the same mating sacrifice on c4 as in the above variation.
31...¥xf4 is a not-so-subtle way to proceed and win, shattering the pawns around White's king: 32.exf4 £xf4 33.¦e7+ ¢f6−⁠+ with a mate in eight, according to Komodo.
32.£d1 ¤c4 33.¦d8! according to the ChessBase article, Sethuraman only had about two minutes left on his clock at this point, with many complications to resolve. 33...¥e7?! the obvious move, which however gives White a lot of breathing room.
33...¤xe3 is flagged by the engine as best, again with the idea of shattering White's protective pawns, although it is hardly easy to calculate. 34.¢xe3 ¥xf4+ 35.¢xf4 ¦g4+ 36.¢e5 (36.¢f3?36...£h5 with a mating net) 36...£f6+ 37.¢d6 £xd8+ winning
34.¦d7µ34...¦ab2 this gives White the ability to get back to equality, with the following move.
34...¤b2+ is probably the simplest line, forcing the win of an exchange. 35.¦xb2 ¦axb2 but Black may have been put off by the following line: 36.¥h4 ¥xh4 37.¦xf7+ ¢xf7−⁠+ with a large advantage, but not so easy to evaluate. For example after 38.£h5+ ¢g7 wins handily, as the Bh4 cannot be captured due to the mate on d2. Calculating all this in time trouble would be difficult if not impossible, though.
35.¥xc4 dxc4+ 36.¢xc4 £e8 (36...e5+!?) 37.¦xb2 ¦xb2 38.£a1 ¦b8 again, what looks like an obvious harmless move turns out to be bad for Black. (38...¦e2 39.¢d3 ¦xe3+ 40.¢xe3 £xd7) 39.£a7+⁠− now Black is the one under major pressure. 39...¢f8 40.¢d3 now White's king is no longer exposed to Black counter-threats. 40...¦a8 41.£b7 ¦b8 42.£h1 playing it safe by preventing ...Qh5 and trading off material, leading into an endgame. (42.£c6!?42...£h5 43.¥d2+⁠−) 42...£xd7 43.£h8+ ¢f7 44.£xb8± queen and minor piece endgames are complicated and difficult, but Shankland manages to convert his two-pawn advantage at his leisure. His opponent could not have been in a positive frame of mind for a long, grinding defense. White keeps threatening to exchange queens while removing his king from annoying checks. 44...£c6 45.£b2 £e4+ 46.¢d2 £g2+ 47.¢c1 £f1 48.¢d1 £d3+ 49.£d2 £c4 50.£e2 £a4+ 51.£c2 £c4 52.¢d2 £f1 53.£d3 £h1 54.£e2 £e4 55.£h2 £b7 56.¢e2 £b2+ 57.¥d2 £b5+ 58.¢f2 ¢g6 59.£g2+ ¢f7 60.£f3 ¥h4+ 61.¢g2 £d3 62.£h5+ ¢f8 63.£d1 ¢g7 64.£g1 an interesting tactical trade of material. 64...£xd2+ 65.¢h3+ ¢f8?
65...¥g5 is found by the engine. It is very counter-intuitive, but having the bishop choose to sacrifice itself on g5 appears to make it impossible for White to make progress after Black regains one of the pawns. For example 66.£xg5+ ¢f7 67.£h5+ ¢g7 White has no checks now and must lose either the e- or c-pawn.
66.¢xh4+⁠−66...£xc3 in contrast with the above variation, White now can penetrate the kingside and threaten Black's king. 67.¢h5 £c6 68.¢h6 £f3 69.£g7+ ¢e8 70.£e5 ¢d7 71.¢g7 £g4+ 72.¢f8 £h4 73.£g7+ starting the final sequence. 73...¢d6 74.¢e8 £h5+ 75.£f7 calculating the won K+P endgame for White. 75...¢d5 Black hangs the queen and resigns, although he was lost anyway.
75...£xf7+ 76.¢xf7 ¢d5
76...¢d7 77.d5 an instructive temporary pawn sacrifice in the ending, undermining the f-pawn and allowing the win. 77...exd5 78.¢f6 and wins.
77.¢e7 and the e-pawn is doomed.
Powered by Aquarium

04 December 2016

Commentary: 2016 Russian Women's Team Championship Round 1 (Lagno - Goryachkina)

The following commentary game (Lagno - Goryachkina, from May's Russian Women's Team Championship) is the first of the last series on this blog of such games in 2016.  It features what can be a very annoying White choice against the Caro-Kann (by transposition).  Original ChessBase report and commentary can be found here.

For me the game has several standout lessons for training purposes:
  • The trickiness of the variation and Black's need to carefully consider how to neutralize White's early pressure.  Goryachkina's innovative choices in the opening (7...g6 and 8...Qd6) required careful calculation up front but paid off in the end.  8...Nb6 also looks like a fine choice for Black, with full compensation for the pawn sacrifice.
  • Black's potential piece activity was evident as of move 11 and by move 20 she was completely dominating her opponent positionally.  All of White's pieces had retreated from Black territory, while Black's knights had established outposts on the other side of the board.  This high level of fluidity in the position was possible due to the lack of central control by White and her underdevelopment, particularly evident regarding the d-pawn and the blocked-in dark-square bishop.
  • Black's ability to accurately and fully calculate for the entire game was impressive, including in the above-mentioned sequence after 8...Qd6, but also at turning points such as move 21.  Seeing moves such as 21...Nf4 and their consequences ahead of time is what master-level chess is about.
  • Finding winning moves rather than necessarily "best" moves.  Black's move 23 is a case in point, where the engine evaluation is much stronger after 23...e4, but Black goes with a more humanly understandable path (23...Nc5), playing ...e4 anyway two moves later.

Lagno, Kateryna (2529) - Goryachkina, Aleksandra (2485)

Result: 0-1
Site: Sochi RUS
Date: 2016.05.01
[...] 1.c4 c6 2.e4 now we have a Caro-Kann 2...d5 virtually the only response that makes sense after 1...c6, although I suppose one could transpose eventually into a Modern Defense or the like without ...d5. 3.exd5 cxd5 4.cxd5 this keeps the opening in its own unique variation.
4.d4 is another transpositional alternative, this time to the Panov-Botvinnik Attack.
4...£xd5 is the main alternative, but White scores 68 percent in the database afterwards. The Black queen will inevitably lose time relocating from d5.
5.¤c3 ¤xd5 6.¤f3 ¤c6 7.¥b5 all natural developing moves by White so far. 7...g6 although not used very often, this variation scores far better than its more classical counterparts, ...Nxc3 and ...e6. 8.£a4 £d6 an interesting choice that requires careful evaluation of the next sequence.
8...¤b6!? is almost always played here. 9.¥xc6+ bxc6 10.£xc6+ ¥d7 11.£e4 ¥g7 almost all of the database games from this point end in a draw, with Black's compensation for the pawn including the two (outstanding) bishops, play against the isolated d-pawn, and good avenues for the rooks.
(8...¥d7??9.¤xd5+⁠−) 9.¤e4 now we are in new opening territory. 9...£e6 10.¤fg5 £d7 11.¤c5 £c7 at the end of the forcing sequence, White has kicked around the Black queen, but Black is not really behind in development, as she will have an easy time getting her bishops out, compared to the Bc1. White's minor pieces are all forward deployed, but not working together particularly well. Komodo assesses the position as equal, but White is the one who can misstep more easily here. 12.O-O
12.£d4!? is the engine's recommendation. 12...¤f6 13.£c4 e6
12...¥g7³ Black is now starting to look more dangerous, as the Bg7 is now a monster on the long diagonal and White has no real threats. 13.¥c4 £d8 the best way of maintaining the tension in the center, not afraid of the following sequence. 14.¤xb7 ¥xb7 15.£b5 O-O a cold-bloodedly correct move. 16.£xb7 ¤db4 eyeing the c2 square and restricting the White queen. 17.£b5 ¦b8 18.£a4 ¤e5 for the cost of the sacrificed b-pawn, Black has far more piece activity, while for White the Bc1 and Ra1 are not playing. 19.¥e2 this is too passive.
19.d3!? would give back material in order to get the Bc1 and Rf1 into the game. 19...¤bxd3 20.¦d1 ¤xc4 21.£xc4 ¤e5³
19...¤ed3µ (19...¤bd3!? also looks good.) 20.¤f3 it is remarkable to compare this position with the one on move 11, as all of White's pieces have retreated while Black's have advanced, and now White is behind in development. 20...e5 Black has an excellent position, but it's not clear exactly what plan is best. Dominating the c-file looks good, while taking the b2 pawn at this point does not. In the game, Goryachkina with this move chose to occupy the enter with the e-pawn. She must have also calculated the next sequence as part of it, perhaps even playing the text move to encourage her opponent to challenge the Nb4. (20...¦c8!?) (20...¤xb2?!21.¥xb2 ¥xb2 22.¦ab1 ¤d5 23.£xa7) 21.a3?! White must have been feeling a little desperate by this point.
21.¥xd3 would have helped White gain some maneuvering room and eliminated one of the two forward-deployed knights, at the cost of a pawn. 21...¤xd3 22.¤e1 ¤xb2³
21...¤f4 a forced move for Black in response, creating a counter-threat against the Be2 while the Nb4 is hanging. 22.¥d1 ¤bd3 now the Nb4 has a place to go and Black is even more dominant. White has no good moves available, although the engine suggests Ne1 as the best defense. 23.g3 ¤c5 a "good enough" type of move that preserves Black's advantage.
23...e4 is what the engine prefers. It would take advantage of the e-pawn's position and launch a decisive attack. For example 24.gxf4 exf3 25.¥xf3 ¦e8 and the f-pawn will eventually fall while Black remains dominant positionally. However, this requires a number of moves to fully unfold and in practical terms it does not look easy to clearly evaluate the situation at the board.
24.£c4 ¤fd3 25.¥c2 e4!−⁠+ now the pawn advances to good effect, sacrificing itself to achieve Black's complete piece dominance. 26.¥xd3 ¤xd3 27.£xe4 ¦e8 28.£a4 this immediately lets Black's queen into d5, but White has severe problems in any case.
28.£c4 ¦c8 29.£b3 ¦e4 and now the rook can transfer to the c-file and pressure the trapped Bc1.
28...£d5 29.¤h4 (29.¢g2 ¦e1−⁠+) (29.£d1 ¥xb2 30.¥xb2 ¤xb2 31.£c2 ¦ec8 32.£b1 £xf3−⁠+) 29...¤xf2 and Black can follow up with ...Re1 and/or ...Bd4 to end the game.
Powered by Aquarium

03 December 2016

An improved version of the Fajarowicz Gambit?

As part of my current improvement plan, I'm (slowly) working my way through Mastering Opening Strategy by GM Johan Hellsten.  One of the exercise games (#43) in the chapter "The Nature of Development" intriguingly reminded me of an improved version of the Fajarowicz Gambit.  For those not familiar with it, it is an audacious variation of the Budapest Gambit that, unfortunately, is also not quite sound.  Below are a couple of the critical lines stemming from the initial gambit, where Black, instead of 3...Ng4 as in the Budapest, plays 3...Ne4 (evaluations by Komodo 10):

Fajarowicz Gambit

[...] 1.d4 ¤f6 2.c4 e5 3.dxe5 ¤e4 4.a3
4.¤f3 ¥b4+ 5.¤bd2 d6 6.exd6 £xd6 7.a3 ¥xd2+ 8.¤xd2 ¤c5 9.¤f3 £xd1+ 10.¢xd1 ¤b3 11.¦a2 ¥e6 12.e4 ¤c5²
4...¤c6 5.¤f3 d6 6.£c2 d5 7.e3 ¥g4 8.cxd5 £xd5 9.¥c4 £a5+ 10.b4 ¥xb4+ 11.axb4 £xa1 12.£xe4 ¥xf3 13.gxf3 £xe5±
Powered by Aquarium

While the Fajarowicz is a fun gambit for Black to study - I went through Tim Harding's book The Fighting Fajarowicz with great interest - ultimately it doesn't work out as well as Black would like, unless White cooperates by not playing the main lines with 4. Nf3 or 4. a3.  It's not necessarily a loser for Black, but with some rather simple White play, Black's otherwise fascinating tactical possibilities and initiative can be neutralized, which are really the only reasons to play the gambit.  I have to give Harding a lot of credit for not over-selling Black's prospects and providing valuable, candid analysis in the book.  Harding also took another look at the opening after the book was published, if you are interested in his commentary.  (Of course he's not the only writer on the Fajarowicz, you can look up others. The short version would be the Wikipedia article, the long version the Budapest Fajarowicz (A51) webliography posted at the Kenilworthian blog.)

Having somewhat regretfully put away the Fajarowicz as a possible weapon in my opening repertoire, I was surprised and a little fascinated by the following game from Hellsten's book.  It is classified as ECO E37 - Nimzo-Indian Classical, Noa Variation.

Bareev, Evgeny (2675) - Ivanchuk, Vassily (2695)

Result: 0-1
Site: Novgorod
Date: 1994
[...] 1.d4 ¤f6 2.c4 e6 3.¤c3 ¥b4 4.£c2 d5 5.a3 ¥xc3+ 6.£xc3 ¤e4 7.£c2 c5 8.dxc5 ¤c6 9.¤f3 £a5+ 10.¤d2 ¤d4 11.£d3 e5 12.b4 £a4 13.¦a2 ¤xd2 14.¦xd2 ¥f5 15.£e3 O-O-O 16.g4 £c2 17.¦xd4 exd4 18.£d2 £xd2+ 19.¥xd2 ¥e4 20.f3 ¥g6 21.cxd5 ¦xd5 22.¥g2 f6 23.¢f2 h5 24.¥f4 ¥c2 25.h4 ¦e8 26.¦c1 ¥a4 27.gxh5 ¦xh5 28.¥g3 ¦e3 29.¦c4 ¦d5 30.¥d6 ¦c3 31.f4 ¦xd6
Powered by Aquarium

The key gambit characteristics for Black arise from his 6th and 7th move choices. With the first, the "Fajarowicz" knight appears on e4 and with the second, Black looks to undermine the White center for quick development.  If you look at the position on move 8, it seems like a classic Fajarowicz structure, with the benefit of White not having any minor pieces developed (just the Queen on c2, which has already been kicked once from c3).  By move 14, the thematic ...Bf5 tactical motif in the Fajarowicz has appeared, with the idea that Black's minor pieces are playing in the center, targeting key squares in White's camp and White's queen.  Could it be that this Nimzo-Indian variation is actually an improved version of the Fajarowicz?  Something to think about for both Fajarowicz fans and players who want a rock-solid opening that still has gambit possibilities.

It's fascinating to see some of these ideas for Black appear across different openings and at high levels, which reinforces several different training ideas for improving players:
  • Studying and annotating master games
  • Varying your opening study and looking outside your current repertoire for ideas
  • Studying everything - nothing you do is wasted time, if you approach the material with a critical eye and look to better understand chess principles and patterns. I doubt I'll ever actually play the Fajarowicz, but having studied the opening I can now recognize key themes about development advantages resulting from gambits, along with particular tactical ideas for Black in related structures.