there is also a threat to bring the knight to e4.

This allows the f-file, on which the king stands, to be opened, which should not have been allowed in any circumstances. Consequently, a more tenacious defence was 22...exf5, after which the strongest resort is 23.exd6 23...Qxd6 24.Rd1 Qc7 25.Nd5 – the white attack will henceforth develop almost of its own accord.

More tenacious was 28...Kg7, after which the cleanest win is 29.Rexf6 Qxf6 30.Qd3 Rd2 31.Qg3+ Qg5 32.Rf7+ Kh8 33.Qc3+ Qe5 34.Qxd2 Qxh5 35.Rxb7.

Black resigned.

The b5-pawn hangs with check, and castling is impossible.

Of course, the cautious 22...Qc4 was also possible, but the developing move chosen by Black is stronger, as it connects the rooks, and keeps the possibility of ...Qc7-c4 in reserve. Of course, Black did not forget about the question of his king’s safety in the centre – the move is possible because of the fixed pawn structure. In addition, Black has already exchanged White’s knights – with them still on the board, there would be tactical motifs.

Like Black, White also needs to refrain from castling, because of the loss of the h2-pawn. He has to go in for exchanges, which are unfavourable – after all, White has an isolated pawn, and his queenside pawns are fixed on dark squares.

An endgame cannot be avoided, so the king may as well stay in the centre, even though he could have castled. White is
concentrating on holding the draw.

It is useful to rule out Black’s knight exploiting the f4-square. In addition, the light-squared bishop is without an opponent, and his pawns therefore belong on dark squares.

White intends to jettison a pawn, but eliminate his weaknesses, and rely on his well-placed pieces to hold the draw. The position after 26.Kxd3 would be passive – defensible, but unpleasant.