LPSupi vs Achuvettan: game review

15…Nxe3! The only good move here: every alternative is clearly worse.

• Takes the bishop on e3.
• Shores up the d5 pawn, which was under as much attack as it had defence (1 against 1).
• The next best move, 15…Nh4, would drop Black's winning chances from 50% to 30% (+2.29).
• 15…Nxe3 trades off the attacked knight instead of losing it to gxf5.
• White was threatening gxf5, winning a knight. The knight takes something with it instead.
• Trades knight for bishop.

16…g6 is a mistake: Black's winning chances go from 47% to 36%.

• White answers 17.Nf6+, which gives check, so the reply is forced.
• Better was 16…O-O (+0.35), which castles: the king gets out of the centre and the rook joins the game.

Better: 16…O-O stops Nxg7+. After 17.Rf3 Bg5 18.Qe1 f6 19.exf6 Bxf6 20.Nxf6+ Rxf6 21.Rxf6 gxf6 22.Qf2, the rooks are traded (+0.35).

18.Rxf7 is a mistake: White's winning chances go from 60% to 48%.

• Black answers 18…Raf8: the rook takes the open f-file, aiming at the rook on f7.
• Better was 18.e4 (+1.06): a pawn break: it challenges the d5 pawn to open lines.

Better: 18.e4 wins a pawn, picking it up with 19.exd5, and threatens exd5, winning a pawn. After 18…Raf8 19.exd5 exd5 20.Nxd5 Rhg8 21.Nf6+ Bxf6 22.exf6 Kc8, White comes out a pawn up (+1.06).

20…Qc4 is a mistake: Black's winning chances go from 51% to 41%.

• White answers 21.Nxe6, which takes the pawn on e6.
• After Nxe6, White wins 1 point of material within 7 moves.
• Better was 20…dxe4 (−0.14), which takes the pawn on e4.

Better: 20…dxe4 creates a passed pawn on e4 and trades pawns. After 21.Nxe4 Qb5 22.Nc5+ Bxc5 23.dxc5+ Kc8 24.Nxe6 Rf7, Black is a pawn down (−0.14).

21.Nxe6!! A real sacrifice: White gives up material and the engine still rates it best.

• Puts the knight on e6 where it can be taken.
• Black can take, but after 21…Kxe6 22.exd5+ Kd7 23.dxc6+ the engine rates it +1.00.
• Takes the pawn on e6.
• Creates a passed pawn on e5: no enemy pawn can stop it any more.
• 21.Nxe6 wins a pawn and stops …Nxd4.
• Takes the e6 pawn and wins a pawn once the captures settle.

• The idea: 21.Nxe6!! offers the knight on e6. Taking it gives White the material back and more.
• What White gives up: 21.Nxe6 puts the knight on e6 where it can be taken by the king. That's a knight on offer.
• If Black takes the knight: 21…Kxe6 is met by 22.exd5+, taking the d5 pawn with check. Then 23.dxc6+ gives check, and 25.d5 creates a passed pawn on d5. Material is level and White is a little better (+1.14). This is Black's best try.
• Without the sacrifice: The best move without the sacrifice, 21.exd5, is clearly weaker (−2.07 against +1.00). Black answers 21…exd5, taking the d5 pawn. That would leave White clearly worse.

26.Kh3 is a blunder: White's winning chances go from 48% to 8%.

• Black answers 26…Qf4, which attacks the e5 pawn, which is undefended.
• Better was 26.Kh1 (−0.21): the king steps out of the firing line (fewer enemy pieces aim at it now).

Better: 26.Kh1 steps the king out of the firing line. After 26…Ke7 27.dxc6 Qxc6+ 28.Qd5 Qxd5+ 29.Nxd5+ Ke6, the queens come off (−0.21).

26…Qf4!! A real sacrifice: Black gives up material and the engine still rates it best.

• Leaves the bishop on h4 where it can be taken.
• Taking is too dangerous: White's best reply is 27.Qg1 Ke8 28.dxc6 Be7 instead (the engine rates it −6.45).
• Attacks the e5 pawn, which is undefended.
• Lines up queen and rook on the f-file, so the two pieces push in the same direction.
• 26…Qf4 wins the queen for 6 points of material, picking it up with 29…Qxf2, and threatens …Qxh2#, with mate in 1.
• Sets up a sequence and wins the queen for 6 points of material once the captures settle, picking it up with 29…Qxf2.

• The idea: 26…Qf4!! leaves the bishop on h4 hanging. Taking it leads to mate.
• What Black gives up: The bishop on h4 is attacked by the king, and 26…Qf4 doesn't save it. That's a bishop on offer.
• What it does instead: It threatens …Qxh2# and mate.
• If White takes the bishop: 27.Kxh4 is met by 27…Rxh2#, and it's mate.
• White's best try: White's best is to leave the bishop alone and play 27.Qg1. Black answers 27…Ke8, steping the king out of the firing line. Then 28…Be7 threatens mate, and 29…Qxf2 takes the queen on f2. Black ends up a piece and a pawn ahead: Black is winning (−6.45).
• Why not save the bishop first?: Saving it with 26…Bg5, the best way to do that, is clearly weaker (+0.42 against −6.45). White answers 27.dxc6+, taking the c6 pawn with check. That would leave Black level.

28.Qh1 is a mistake: White's winning chances go from 7% to 0%. It allows mate in 3.

• Black answers 28…Bd8, which brings the bishop, Black's least active piece, into play (from 3 to 6 usable squares).
• After Bd8, Black wins 1 point of material within 4 moves.
• Better was 28.Qg1 (−6.89): shores up the h2 pawn, which was under as much attack as it had defence (2 against 1).

Better: 28.Qg1 threatens c7, winning a bishop and a pawn, picking it up with 29.Kxh4, and stops the mate threat …Rxh2#. After 28…Be7 29.Qxf2 Qxf2 30.Ne4 Qh4+ 31.Kg2 Qxg4+ 32.Ng3, White is the exchange plus a pawn down (−6.89).

• Both sides followed opening theory up to 7.Bd3: 13 moves in all, 7 by White and 6 by Black.
• Theory runs out after 7.Bd3: too few games between strong players reached this position to go further.