Mihajlov vs Bänziger

In the Najdorf with 6...e5 Black has locked the centre and permanently weakened the d5 square in return. The standard answer is 9...Be6, and it is not a question of development but of occupation: the bishop looks at d5 and takes any knight that dares to go there.

The move played was 9...b6 with 10...Bb7, and there the bishop sits on the long diagonal behind its own pawn on e5 and does not cover d5. The difference is not quite a third of a pawn, so it is small, but it is the cause of everything that happens over the next ten moves. White manoeuvred with 10.Nd2, 12.Nc4 and 16.Ne3, and on move seventeen a knight stood on d5.

The rule behind it reaches well beyond the Najdorf: when you weaken a square in the opening, develop the piece that covers that square, not the one that looks best. A bishop on the long diagonal looks active, a bishop on e6 would have been working here.

Black has traded off both white knights on d5, paid for it with a knight and the bishop, and is left with two weaknesses: the backward pawn on d6 and the pawn on b4 that nobody defends. Defend passively here and you will be defending for a long time. 21...d5 solves both in one move, because a backward pawn stops being backward the moment it advances.

The pawn is not really on offer at all. After 22.exd5 e4 23.Bxe4 Bc5+ 24.Kh1 the black pieces come into the game with check and with tempo, and the two pawns are worth less than that: Black is clearly winning there. White declined with 22.Be3 and left the tension standing, and after 23...Bc5 24.Bxc5 Qxc5+ Black had the better game. The alternatives 21...Rb8 and 21...Qa7+ cost a third of a pawn each, and that third goes straight into defending d6.

Pushing the backward pawn is one of the few levers that transforms a passive position at once. Check it in every position where you carry such a weakness, and calculate it even when it costs a pawn at first. Active pieces are nearly always worth more than an intact pawn on a square you have to babysit anyway.

Black is about a pawn and a half better here: the queen on d4 is the best piece on the board, the b4 pawn has turned into an asset, and the white king has nothing but pawns in front of it. The move that converts this is 30...Nh5, heading for f4. From there the knight hits h3 and g2, and the white king has no defender left that covers both squares.

The move played was 30...h5, the pawn instead of the knight, and after 31.b3 h4 a good half of that pawn and a half is left. The difference is who gets to the kingside faster: 30...Nh5 brings a piece into the attacking zone, 30...h5 brings a pawn, and pawns need two moves for what a knight does in one.

When you are clearly better and the enemy king is weak, count the pieces already attacking before you make any pawn move. Fewer than three, and the right move is almost always the one that brings the next piece over. Pawn storms are the tool once the pieces are in place, not before.

The knight now stands on f4, where it belonged ten moves earlier, and the position is worth a good two pawns, provided Black plays 40...e4. The pawn advances, the a1 to h8 diagonal opens for the queen, and the f3 pawn becomes a target. The main line runs 40...e4 41.Rh8 Qd4+ 42.Kh2 Nxg2 43.Qf8+, and there queen and knight sit next to the white king while Black wins the material back.

The move played was 40...Kh7, a safety move that takes the king out of every check and misses the moment for it. After 41.Kh2 Nh5 42.Rd1 the position is level again, and Black has to start over.

This is the pattern I see most often in strong amateurs: you are better, you tidy up prophylactically, and two moves later you notice that the advantage was an opportunity and not a property. An advantage that is not converted into a concrete move disappears. The question is not whether the position is safe, it is which move forces something here and now.

With 44...Rd7 Black put the rook undefended in front of the white rook, and 45.Rxd7 is the only answer a human plays. That was not an oversight, it was a calculation: Black has queen and knight next to the white king, and two pieces are enough for perpetual check when the squares are right.

In the diagram there is exactly one move, and everything else loses. 45...Qc1+ 46.Kf2 Nh1+ 47.Ke2 Ng3+ 48.Kf2 Nh1+, the position repeats, and White cannot get out of the checks. Draw. The alternatives show how narrow the path was: 45...Kg7 46.Rxf7+ Qxf7 47.Bxf7 Kxf7 48.Qc7+ costs the queen, and 45...Kh6 loses just as quickly.

Before 44...Rd7, the move 44...Qe3+ was objectively better and held a good pawn. Fabian Bänziger chose the draw he could calculate instead, and in a team match that is a legitimate choice. What matters is the order: calculate the perpetual to the end first, then give up the material. Do it the other way round and you sacrifice, then go looking for the checks.