Bänziger vs Ivic

From a Caro-Kann Advance a position has arisen in which White holds one thing and everything else arranges itself around it: the d6 square. His own pawn on e5 covers it, the black pawn on c6 does not interfere, and the knight on e4 can go there in one move. That is not a detail, that is the core of the position, and the strongest move is the one that uses it: 15.Nd6, hitting f7 and b7 at once.

After 15.Nd6 Bxf3 16.Qxf3 Qd7 17.Qh5 g6 the knight on d6 is not a piece you get rid of again, and White stands a good half pawn better. The move played was 15.Rc1, and it is not a catastrophe, but it gives back half of the advantage: four fifths of a pawn become a bare third. Fabian Bänziger spent sixteen minutes and twenty-five seconds on this move, the longest think of the game.

The rule behind it fits into one line: when the outpost is there and the piece is ready, occupy first and organise second, not the other way around. A rook on c1 is the preparation of a move that forces nothing without the occupation of d6. The knight on d6 is the move the preparation is supposed to prepare.

The position is level. Two black knights on d5 and f5, a white central pawn on e5, neither side has more than a quarter of a pawn in hand. The one break that forces anything is 31.g4, and it is concrete: the pawn attacks the knight on f5, and after 31.g4 Ndxe3 32.fxe3 Ng3 33.Nd2 Nxe4 White keeps the balance and dries out the position.

The move played was 31.h5, with one minute fifty-eight left after the move. The h-pawn attacks nothing. The answer is the move White no longer had the time to calculate: 31...c5, and with it the black centre breaks open. 32.dxc5 Ndxe3 33.fxe3 Rd3, and the position has swung from zero to a good half pawn in Black's favour.

Objectively that was a tactical error, not a move in the wrong order. From a human point of view it was the move you make in time trouble because it does not seem to require calculating any of the opponent's replies. The rule sits exactly there: in time trouble you fix your eye on the move that touches the opponent's piece, not on the one that pushes a pawn of your own. A pawn move on the edge changes the position by one square; the pawn in the centre that you let attack you changes it in every square.

From move thirty-four onwards Fabian Bänziger did not have a full minute on the clock, and between move thirty-four and move forty he never once climbed back above two. In sequences like this it is preparation that decides, not calculation: 33.fxe3 was the only move that still held the position, 34.Qc4 the only one that defends against Qb3, 35.Qxb3 the trade that removes the queen from the board before it does anything, 36.Rd2 to cover the second rank. All played in seconds, all the best move.

The one move in the chain that the clock really contaminated is 38.Nf3. The knight retreats, instead of the 38.Rc1 that targets the knight on d5: 38.Rc1 Rg5 39.Nf3 Rxh5 40.Rxd5, and the position is only a good tenth of a pawn worse. After 38.Nf3 Rxc5 39.Re4 f5 40.exf6 Nxf6 White slips to a good four tenths of a pawn behind, and that is the value the game takes into the ending.

The rule for your own clock is a familiar one: in time trouble you play the moves you have calculated already, and stop trying to calculate the ones you would like to play now. Fabian Bänziger reached move forty with one minute and twelve seconds, more seconds than the run of moves between should have allowed. That is a craft in itself.

A rook and knight ending against rook and knight, three pawns against three on the kingside, all in one column. The material is unremarkable, the verdict is level-headed: for Black it is a good tenth of a pawn, less than the width of a square. A grandmaster with twenty-two minutes and White with thirteen plays such a position out for thirty moves. The draw has to be defended, not offered.

The white moves read like a manual on holding: 50.Ne5 puts the knight active, 51.Rd6 ties the e6 pawn to its defender, 53.Re7+ takes the seventh rank, 54.Rf7 pins the knight on f6. No move is more than a fraction off the best, no sequence hands a grandmaster anything. When the white clock falls under five minutes a second time from move fifty-six, the defence is already standing.

The rule for endings with equal material and pawns on one side is the quiet neighbour of the whole endgame theory: without a passed pawn and without a weak king, there is barely a way to win a thing like this. The defender does not need to outdo the activity of his pieces, he needs to hold it. Fabian Bänziger did that, and the repetition 68...Nd4+ 69.Kd3 Ne6 70.Ke2 together with the closing 72.Kxg3 Nf5+ 73.Kh2 Rxg2+ was the logical conclusion.