Hryzlova vs Herfesa

10...Nf8 sets the knight far away from the play, because 10...g6 first would have been safer before 11.Nf5. White stood clearly better after this move, and Sofiia Hryzlova used the time to turn the clear advantage into a concrete plan: nearly thirty-two minutes for 11.Nf5.

The move asks Black two questions at once. The g7 pawn without the f8 knight is guarded only by the king, and the e7 bishop is under attack too. 11...Ne6 answers both together, because the g7 square is now covered by the knight and the d8 queen keeps defending e7. White declined the immediate trade, because Nxe7+ Qxe7 collects the bishop for the knight, but is worth more a few moves later: 12.Bb2 Rb8 13.Rc1 O-O 14.Rc2 g6 15.Nxe7+ Qxe7, and the pieces are ready so that the black kingside stays airy for good.

The lesson of this passage: when you already stand clearly better, the move to play is not the one that looks quickest but the one that fastens the opening lead down for good. The thirty-two minutes here were well spent; they paid off four moves later with the bishop trade.

The bishop trade was due, 15.Nxe7+ Qxe7 came, and the dark squares belonged to White. Objectively the position at the transition to the middlegame was clearly better, and the plan should have been simple: 16.e3, prop up the d4 pawn, then Nc3 to e2 and the heavy pieces along the c-file. What came was 16.Na4?!, eyeing c5.

Objectively that is a tactical error, because after 16...cxd4 17.Bh3 Ng5 18.Bg2 Black holds a clean extra pawn, and the white opening edge is gone. There were eighteen minutes and change left on the clock after nearly ten minutes of thought, and above the board sat a move the opponent on the next board might not find but which two players in three at this level do choose. Practically 16.Na4 was the human try, part of the same queenside manoeuvre. Objectively it was a pawn.

The move at the other end of the board, 16...cxd4, did not come. Instead 16...Rfc8?, a rook move meant to occupy the c-file, and with it the last good answer. Now 17.dxc5 would have held on to a full pawn and cleared the position; 17.Ba3?! chose the third rank instead, and within two moves White had turned a clear edge into an equal game without the clock forcing her to do it.

A knight move to the rim, against which the opponent has a concrete answer, is not less risky because they do not find it. The next opponent will.

Six moves before the time control, with 9:40 on the clock and an idea in her head: 24.f5 opened the f-file, and the g6 square lit up. 24...Nd8?? took the knight off the centre instead of 24...gxf5 to close the line. The whole game turned in a half-move. White was winning.

The winning move is a single one: 25.fxg6 fxe5 26.gxh7+ Qxh7 27.Nxd5, and the black king stands there without a pawn cover, while White collects the d5 pawn and brings the heavy piece to the fourth rank with check. No brilliancies are required, only counting to the last capture. Given a good hour, most players at any level play 25.fxg6 straight out of the hand.

Instead came 25.Bf4??, a move that pulls the bishop out of the centre into the corner and gives up the whole kingside idea. 25...g5 asked the bishop the question at once, and 26.Bc1 put it back on the square it had stood on at move thirteen. A winning attack had become a small edge, and the clock, which read 14:44 before 25.Bf4, was now working against White.

A winning position is not the win, it is an obligation to find the concrete sequence. When the enemy king has no pawn cover and yours still does, the first place to look is captures and checks, not the redeployment of your own pieces.

Between moves 25 and 30 Black handled the position worse than White: 26...Nd6? and 29...Qf7?! gave Sofiia Hryzlova the reins back, and 30.Qg4+ Kh8 tucked the king into the corner with the wedged f6 pawn as the tip of the attack. The clocks read 3:42 for White and 3:31 for Black. One move would have held the attack together.

31.Bh6 claims the diagonal towards g7, with ideas such as Qh4 and Bxg7+ or the rook to f3 into the attack. After 31.Bh6 Ne8 32.Qh4 White holds the whole kingside attack together, and the black king has no square left that is not covered. Objectively that was the move, and practically it was also the prettiest.

What was played was 31.Re2?, a rook move that lifts the rook to the second rank but abandons the attack. 31...Ne4 came at once, the reply that fitted any rook move, because the black d6 knight had been waiting seven moves for the jump to e4. After 32.Bxe4 dxe4 the attacking diagonal was gone, and the c3 pawn stayed alive. The win was no longer on the board; the draw still was.

In time trouble, the move to make is the one that shifts the piece that is doing the most work, not the piece that is doing the least. The e5 bishop was holding the whole attack together here; the f1 rook was only standing.

After 32...dxe4 both sides had to keep their nerve in mutual time trouble. 33.Bf4? gave a little back, because 33.d5 would have opened the diagonal towards the king and brought the white queen into the picture. 33...Nb7?? and 34.Be5?? passed as small errors because they cancelled each other. Then came 34...Bb5??: with forty-three seconds on the clock, the bishop attacked the e2 rook, and the position was clearly winning for White. The white clock showed 2:08.

The move the position asked for was 35.Rh2, taking the rook out of the pin and keeping the pressure on f7. Instead came 35.Qxe4?? after a minute’s thought, with a minute and a second left; the move captures a pawn but does not hold the rook, and 35...Bxe2 took the exchange. Right after, with 1:19 on the clock, came 36.Qxb7??, a second pawn-grab on the wrong side of the board. 36.Qxe2 would have won the exchange back and kept the game somehow alive. The b7 pawn cost the knight and cut the attack in half.

What these two moves teach applies to any time-trouble game: a pawn that can be taken is not the same as a pawn that should be taken. When you capture in seconds, the first thing to see is the reply, not your own move.

After move 40 the clock came back, but the position was decided. A rook for a knight, a passed pawn on c2, the black king safe and the white king pinned to the f6 pawn nobody could still defend. 45.Kh2 Rd1 nailed the f1 rook down, and Sofiia Hryzlova gave the game up.