Commentary: U.S. Women's Championship 2021, Round 3 (Abrahamyan - Zatonskih)

Following two previous losses in rounds 1 and 2 (an "0-0" result, also known as "castling short") in the 2021 U.S. Women's Championship, WGM Tatev Abrahamyan in the third round scored for the first time against one of the strongest players in the field, veteran IM Anna Zatonskih. While Abrahamyan must have been desperate for a win at this point, as White she did not choose an aggressive opening, instead going into a positional double-fianchetto Reti. 

Some strong players have followed this kind of strategy for a long time - see Kasparov's must-win victory in the English over Karpov - and it is also a very current approach, thanks to its successful use by world champion GM Magnus Carlsen. The idea is to reach positions where you can "just play chess" and outplay your opponent, rather than trying to overpower them with deep preparation or tactical wizardry; at the top levels, that is rarely possible, in any event. In this particular case, Abrahamyan may also have wanted a calmer approach to start, given the grueling previous two rounds.

Abrahamyan's first aggressive choice comes on move 12, when she chooses a disruptive move in the center to directly challenge Black's forces. Instead of closing the center in response, Zatonskih exchanges on e4 and gives White a freer game, which is evidently what Abrahamyan wanted. However, White's pieces start getting tangled and by move 22 Black has a nice position. However, Black does not have a knockout plan and decides on move 23 to avoid a bishop exchange, which is significant strategic decision. White manages to untangle her pieces, then take the initiative by planting a knight on d5 - important psychologically, even if in an objectively equal position. Although White chooses to exchange the knight, d5 is later occupied by a rook and the exchange results in a winning position, with a final tactical flourish.

Credit goes to Abrahamyan for not buckling under Black's small but real positional pressure and advantage - illustrated at the bottom by the the HIARCS Chess Explorer Pro "Evaluation Explorer" chart - while finding ways to make herself more active and eventually target Black's weaknesses (for example the a-pawn). It's also worth noting the game highlights a number of themes mentioned in "The fundamental importance of the relative value of pieces" post, especially regarding piece exchanges.

[Event "U.S. Women's Chess Championship 2021"] [Site "http://www.chessbomb.com"] [Date "2021.10.08"] [Round "03"] [White "Abrahamyan, Tatev"] [Black "Zatonskih, Anna"] [Result "1-0"] [WhiteElo "2344"] [BlackElo "2422"] [EventDate "????.??.??"] [ECO "A11"] [PlyCount "93"] [Annotator "ChessAdmin/Dragon by Komodo 2.6.1"] [BlackClock "0:15:00"] [BlackFideId "14101572"] [TimeControl "5400+30"] [WhiteClock "0:11:47"] [WhiteFideId "13301918"] 1.Nf3 Nf6 2.g3 {going into the King's Indian Attack formation.} 2...d5 3.Bg2 g6 {Black chooses a more symmetrical formation for her bishop development, although she has already staked out territory in the center with ...d5.} 4.b3 {choosing a double fianchetto approach and waiting to challenge in the center.} 4...Bg7 5.Bb2 O-O 6.O-O a5 $5 {this is very commital with the a-pawn and of course does nothing directly for Black's development. By following the usual rule of thumb by reacting in the center to a wing advance, White may gain a slight advantage, or at least an easier game.} 7.c4 c6 {logically reinforcing d5 and blunting the White bishop on the long diagonal. However, it makes ...a5 look even more out of place.} ( 7...dxc4 $6 {this might look attractive at the amateur level to some players, but positionally White would effectively gain a central pawn. Also, the half-open b-file is likely a strength, while White's isolated a-pawn can't effectively be targeted by Black.} 8.bxc4 $14 ) 8.d3 {again choosing a more restrained approach toward development, keeping options open in the center. The main point is that the Nb1 will go to d2 rather than c3.} 8...Re8 {here ...a4 or ...Na6 are much more common in the database, which is to be expected, since they are consistent with the ideas behind the earlier ...a5 advance. However, the text move scores significantly better.} 9.Nbd2 ( 9.Qc2 $5 {is preferred by the engine and scores much better in the database as well, although the text move is more common. White waits to commit the knight and with the queen move clears the d1 square for a rook, protects the Bb2, and exerts more control over e4, while also establishing a presence on the c-file.} ) 9...Nbd7 10.Rc1 {improving the rook and completing White's development. Black now plays the logical follow-up to previous moves, seizing space in the center.} 10...e5 11.cxd5 {this is almost obligatory, as otherwise Black can get an even stronger center. It also opens up the file for the Rc1.} ( 11.e3 {would keep the tension in the center, but Black has more space and better prospects.} ) 11...cxd5 {recapturing with the Nf6 would reduce the central pawn presence and give White a nice knight outpost on c4.} 12.e4 ( 12.Rc2 {would be a more conventional and less commital choice, with a slow buildup on the c-file possible.} ) 12...dxe4 {Black reacts by giving White the type of game she evidently wants, rather than closing the center.} ( 12...d4 {the advanced pawn cramps White, although not decisively after Nc4.} ) 13.dxe4 b5 {taking away the c4 square from White's knight and gaining space.} 14.Qc2 {overprotecting e4 and projecting power down the c-file.} 14...Ba6 {clearing the c8 square for a rook, while also developing to a good diagonal.} 15.h3 {while this takes away the g4 square from the Nf6, this seems like more of a strategic waiting move.} 15...Bh6 {pinning the Nd2 and getting the bishop to a more active diagonal.} 16.Qb1 $6 {this allows Black's queen to develop effectively.} ( 16.Rfd1 $5 {and now if} 16...Qb6 17.Qc7 ) 16...Qb6 17.Rcd1 {White now fully abandons the c-file, which seems unfortunate. However, Black does not react strongly to it.} 17...Nh5 ( 17...Rad8 {would develop the rook to a useful file, collectively making Black's pieces more stronger than White's; for example ..Nc5 would be a strong follow-up.} ) 18.Rfe1 {clearing the square for the knight to be redeveloped.} 18...Rad8 19.Nf1 {White improves her worst piece.} 19...b4 20.Ne3 {an immediate logical follow-up, but this was not necessarily urgent. The engine suggests Qa1 or Rd5 as other possibilities.} 20...Bxe3 {Black profitably exchanges her bad bishop for good knight.} 21.Rxe3 f6 {this logically reinforces e5 and makes White's pressure on the pawn irrelevant, freeing up Black's pieces from protecting it. However, it takes away a maneuver square for the Black knights.} ( 21...Nxg3 $5 {is an interesting tactical possibility. The point is that the knight cannot be captured by the f-pawn, which would leave the Re3 hanging.} 22.Nxe5 Nxe5 23.Rxg3 Rxd1+ 24.Qxd1 f6 $10 ) 22.Nd2 $6 {this avoids the ...Nxg3 threat, since the Re3 now protects the g-pawn, but White's pieces are placed awkwardly and the Nd2 interferes with the Rd1. Black now exploits this.} ( 22.Nh2 ) 22...Nc5 23.Bf1 {needed to cover the d3 square.} 23...Bb7 {an important strategic decision, done to avoid exchanging bishops. Black may have felt that keeping the pressure on was more important. However, e4 is overprotected and now White's knight goes to a powerful square.} ( 23...Bxf1 24.Nxf1 $15 ) 24.Nc4 Qc7 25.Bg2 {logically reinforcing e4, since d3 no longer needs to be covered.} ( 25.g4 $5 {is the engine's suggestion, which leads to much more aggressive and tactical play.} 25...Nf4 26.g5 fxg5 27.Bxe5 Rxd1 28.Qxd1 Rxe5 29.Nxe5 Qxe5 30.Qd8+ Kg7 31.Qxa5 {is one line.} ) 25...Ng7 {a good idea to redeploy the knight, but perhaps a little premature.} ( 25...Rxd1+ {this is a recurring idea, swapping rooks and getting Black's remaining rook to a good post on d8.} 26.Qxd1 Rd8 27.Qc2 Ng7 ) 26.Ree1 {White now has managed to untangle her pieces and is at parity with her opponent. Black now must try to find a useful plan.} 26...Nge6 27.Ne3 {although the knight looks good on c4, it will be even better on d5. Black has to think about how to cover f6 now.} 27...Qf7 $6 {this unnecessarily defensive move seriously reduces the scope of Black's queen.} ( 27...Ng5 {the engine's alternative, but certainly not an obvious one for a human. The point being} 28.Nd5 Qf7 {and now White has to protect the e-pawn with f3, otherwise for example} 29.h4 ( 29.f3 Bxd5 30.exd5 Rxd5 {is at least equal for Black.} ) 29...Ngxe4 30.Bxe4 Nxe4 31.Qxe4 Bxd5 $19 ) 28.Qc2 Ba6 $6 {this actually moves the bishop away from the most important action.} ( 28...Rc8 ) 29.Nd5 $14 {White may only have a slight plus objectively, but now she has the initiative for the first time.} 29...Rc8 30.Qb1 {getting the queen away from the line of fire of the Rc8.} 30...Nc7 {Black needs to challenge the dominant Nd5.} 31.Nxc7 {here the engine assesses that it's better for White to maintain the knight on d5, but she may well have worried about maintaining an isolated d-pawn after an exchange.} ( 31.h4 Nxd5 32.exd5 Red8 {leads to complicated play, with the engine showing a slight plus for White.} ) 31...Rxc7 32.Bc1 {looking to redeploy to a more active diagonal.} 32...Ne6 $6 {this in fact appears to be a less effective square for the knight.} ( 32...a4 {is the active plan suggested by the engine.} 33.bxa4 Nxa4 {threatening ...Nc3.} ) 33.Be3 Rd7 34.Rd5 {once again White benefits by seizing the d5 square, this time evaluating that a trade will be quite beneficial. Black fails to see this, however.} 34...Rxd5 $6 ( 34...Nd4 {played first is the engine's solution.} 35.Bxd4 exd4 36.Rxa5 Qe6 {and Black is down a pawn, but has compensation in the form of the mobile passed d-pawn and active pieces.} ) 35.exd5 Nd4 36.Qd1 {simply threatening to win a pawn by exchanging on d4.} 36...Nb5 37.Bb6 $16 {and now White targets the weakness created by Black's advanced a-pawn.} 37...Nc3 38.Qd2 Bb7 {unfortunately for Black, the passed d-pawn is mobile and White controls the queening square.} 39.d6 Bxg2 40.d7 $1 ( 40.Kxg2 $4 Qb7+ $19 ) 40...Bxh3 41.dxe8=Q+ ( 41.d8=Q $5 Rxd8 42.Qxd8+ Kg7 43.Bxa5 Qb7 44.Qc7+ $16 ) 41...Qxe8 42.Kh2 Be6 43.Bxa5 {now Black simply does not have compensation for the material, as White's rook will be quite effective in the endgame. As often occurs with a major shift in game evaluation, the player on the downward trend collapses.} 43...Bd5 $2 {now White has a tactical finish, due to the inadequate protection of the bishop.} ( 43...g5 {and the best White has is to exchange queens and grind it out.} 44.Qd8 Qxd8 45.Bxd8 Kf7 46.Ra1 ) 44.Bxb4 g5 {hoping to trick White with a false sacrifice. The point of the text move is that it now gives Black the possibility of ...Qh5.} 45.g4 $1 {this covers h5 and ends Black's tactical hopes.} 45...Qg6 46.Qxc3 Qh6+ 47.Qh3 1-0