Saydaliev vs Bogner

Seven minutes fourteen seconds Sebastian Bogner thought here, and then he played the move that looks like nothing: 12...Nb8, the knight back to the square it came from. The reason lies behind it: the bishop on b7 looks across half the board at the white king, and a knight on d7 of its own takes part of that diagonal away. The knight belongs on c6, and the only road there runs through b8.

The alternative was 12...h6, a waiting move that costs a third of a pawn, or 12...Qb6 for the same price. Both leave the knight where it obstructs. The retreat is the best move, with a clear margin over everything else on offer.

After 13.a4 Nc6 the knight stood where it belonged: eyeing d4, eyeing e5, with b4 and a5 behind it. From here Black built the position up in five quiet moves, and by move sixteen the small advantage the opening theory grants White had turned into a small advantage for Black. The rule: a square where a piece does nothing is the wrong square, even when the right one looks two moves away. Check the detour before you rule it out.

White had played 16.h4 on the previous move and started a pawn storm without having the pieces to back it up. The black bishop on b7 already looked at e4 and g2, the black queen sat on b6, and both rooks had moved to central files on move sixteen. When you begin a king attack with a pawn while the opponent has three units on the attacking file, you spend the tempo you cannot afford.

17.Ne2 is the move you reach for in such moments and the one that does not work here. Its job is to cover f2, the square Black is preparing to hit with ...Ng4. 17.Qe2 does that job, because the queen on e2 defends f2 directly; the knight on e2 does not, because from e2 it does not see f2 at all. Sebastian took fifteen minutes to pick his reply; he had good alternatives to choose between.

What he chose was 17...Ng4, the move that hits f2 and forces the white bishop out to h3. The main line is not a combination but a bundle of small things: 17...Ng4 18.axb5 Nb4 19.Qc3 c4, and Black stands about two pawns better, with pressure on b3, a pawn on c4 that opens the b7 bishop’s diagonal, and a white queen with no good square. White chose instead 18.Bh3, and after 18...h5 he was a good three pawns down.

The rule behind it slipped out from under him in round five: before any wing attack, count the enemy pieces that already stand in your own half. One is enough to delay it, two enough to refute it. The move that vacates the square instead of defending it looks different on paper and has the same effect.

White had traded the h3 bishop for the g4 knight, accepted a doubled g-pawn on g4 for it, and with 20.Ne5 planted the knight aggressively in the middle. From e5 the knight attacks the knight on c6, blocks the d-file, and looks at g6 and f7. Three answers look reasonable here: take the knight, put a knight on b4, or push the pawn.

Sebastian thought eighteen minutes forty-eight seconds. At the end of that time he played 20...f5, the move that twelve percent of players in this position pick and that is nonetheless the best. The pawn attacks the knight, and White cannot take on c6 safely: 21.Nxc6 Qxc6 22.axb5 axb5 23.Be5 fxe4 24.dxe4 Rxd1+ leaves Black a good two pawns up with an open diagonal for the b7 bishop. The alternative 20...Nxe5 holds less, and 20...Nb4 gives a smaller part of the advantage back.

What sets 20...f5 apart is not the one line but the number of them. The pawn changes the structure, opens the diagonal, and threatens ...fxe4. White has to answer three things at once, and none of his pieces has more than one job free. That is why a player of this strength calculates for eighteen minutes and still chooses the move almost nobody chooses: because it answers more questions than the opponent can ask in two moves.

White played 21.a5, the try to prise open the queenside while Black was helping himself in the centre. 21...Nxa5 took the pawn with the knight, the move sixty percent of players find here, and Black was up material. Two moves later the black bishop stood on b7, defending the knight and waiting for the diagonal that was still to come.

Five minutes six seconds after move thirty, under four minutes after move thirty-three, and seven moves left to the control. Black is a knight up, with an attack on the white king and a bishop that supports it. What must not happen now is a move that gives the advantage away; what easily can is a move that extends the game. Sebastian took the second route with a run of queen moves that ask the opponent nothing and give the clock a rest.

After 31...Qe7 32.Qc2 Qe4 33.Qc7+ Qe7, both Black queen moves on moves thirty-two and thirty-three are a shade less precise than the strongest reply on offer. More precise would have been 33...Ke8, walking the king out of the checks and ending the repetition instead of going back to e7. The practical value of the repetition was larger anyway: the opponent had to decline it, every decline cost him time, and the advantage was so clear that a third of a pawn less was no reason to take a risk.

What these moves show, and the clock column of the broadcast confirms, is that the opponent spent no more time than he had to: 31.Qc7+, 32.Qc2, 33.Qc7+ and 34.Qc2 all played in seconds, and he waited for Sebastian to make the first bad move. That is the art of the game played by the higher-rated side: you sacrifice not the pawn but the repetition, and you buy the three minutes that are then enough for the decisive move.

The rule: in a clearly won position with little time, the move that eases the clock is preferable to the move that grows the advantage only when it does not endanger it. Repetitions meet both conditions if the opponent has to decline them, which he did here every time, and that is why they work for two moves until the opponent’s nerve breaks.

Four queen moves the opponent had ridden out in the repetition; the fifth was where it went wrong. 35.Qa2 would have held the position he was already losing, and the rest of the way to the time control besides. 35.Qc3 is the one move that leaves the a1 to h8 diagonal alone and hands the rook the h-file.

35...Rh8 has a single content: Rh1 mate. The bishop on e4 takes h1 and g2 away from the white king, the rook on h1 the rest of the rank. Two answers postpone the mate for a move, and both lose. 36.Ng2 steps between the bishop and h1 so that Rh1+ can be met by Kxh1; the reply is 36...Bxg2 37.Kxg2 Qe4+, and White has no saving move. The second defence is 36.Kf1 Bf3 37.Qxf3 gxf3, which gives up the queen for the bishop and only postpones mate by a few moves. In the game 36.Ng2 Bxg2 appeared on the board, and the opponent resigned while still looking at the capture.

Between the move Sebastian missed on move twenty-nine and the one he found on move thirty-five is the same rook move. After 29.Nf4, Rh8 was already the strongest move on the board and would have ended the game at once; the move played was 29...Qxg5, which took another pawn and cost six more moves. The rule for your own games: when the enemy king stands without cover and an open file lies next to it, the rook to that file is almost always the move to calculate before you win a pawn.