The greatest difficulties arise when your calculations fail to support your intuitive judgement. Today's position, from the fourth round of the London-Peking match, is a good example of the potentially harmful results of such confusion. The diagram position occurred after Chris Ward's 34th move as White, in his game against Peng Xiaomin.

Black must initially have felt content. His knight is well placed and its position may be stabilised by playing g5. The pawns on b2 and a5 are vulnerable to attack by Nc4, and the white bishop is passive.

But when you start to analyse, things become less clear. After 1...g5 2.Rh1 White's rook becomes annoyingly active. If rooks are later exchanged, Black may also be concerned about Bf5 and Bc8, or even (after 2...Rh8 3.Rxh8 Kxh8) f4 and penetration of the white king to e6.

If he is worried about the white rook, Black could play 1...b6 (or 1...b5) 2.axb6 Rb8, when the b2-pawn will always need a defender - but why should Black, who thinks he stands better, exchange the weak white a-pawn?

The seeds of confusion are sown; now watch how the game went: 1...Nc4 (going for the immediate win of a pawn, even if this does mean letting the knight drift away from the centre of the action) 2.g5! Rh8 (2...Nxb2 3.Rb1 is good for White, while 2...Nxa5 3.Rh1 Rh8 4.Rxh8 Kxh8 5.Bxg6 Nc4 is quite wild) 3.Re1! Nxa5 4.Bc2 Nc4 5.Re6 (Black must by now have been wishing he had kept his knight on e5; he is already clearly worse) 5...Rf8 6.Rxg6+ Kf7 7.Re6 Nxb2 8.Rxd6 (Rook, bishop and passed pawn are always an effective attacking combination. Here there are just too many passed pawns.) 8...b5 9.Rxa6 b4 10.Ra7+ Kg8 11.d6 Rd8 12.Ra2 and Black resigned - his knight is lost after 12...Nc4 13.Bb3. A finely played endgame by the British champion, and a model of confusion by his opponent.