Federer vs Rukundo

The Winawer with 3...Bb4 lives off pinning the c3 knight and playing against the doubled pawns after the exchange on c3. 4.Ne2 takes exactly that away: the knight on e2 covers c3, and if Black captures there, White recaptures with the knight and has no weaknesses. A very practical choice, especially against an opponent who may know his theory better than anything else.

What follows is a lesson about time. 5.a3 Ba5 6.b4 Bb6 7.Bb2 c6 8.g3 Bc7 gives White two queenside pawn moves and the long diagonal, while the black bishop moves three times and ends on c7, staring at its own pawn chain. Then 9.Bg2 puts the second bishop on a long diagonal as well.

For your own games: a bishop that keeps retreating in front of a pawn advance costs tempi, and your opponent turns those tempi into space. If you own such a bishop, trade it rather than move it three times.

This is the most important position of the first half. Black has just played 13...Nf6 and placed the knight exactly where a pawn advance hits it. 14.e5 is the move: the knight has to go back to d7, the centre belongs to White, the g2 bishop looks down the long diagonal, and the position is around one and a half pawns better for White.

What was played is 14.Nf4, and afterwards the advantage is gone. After 14...Bxf4 15.Qxf4 Black has traded his worst piece, the bishop on c7, for White's most active knight, and the centre is up for discussion again. The jump to f4 looks active, but a knight standing on a square where it can be captured, with the trade helping the other side, is not an active knight.

The rule behind it holds in every opening: when an enemy piece steps in front of one of your pawns, first check whether the pawn can advance. A push with tempo is almost always stronger than a piece move that only looks active.

Here Black could have released the tension with 15...dxe4 and come close to equality. Instead came 15...Bc4, a move that looks active and puts the bishop on a square where it guards nothing. This time Colin takes his chance: 16.e5 is the best move by a clear margin, everything else is about a pawn worse.

After 16.e5 Nd7 White has the space advantage he passed up two moves earlier: the pawns on d4 and e5 take f6 and g5 from the black knight, and the b2 bishop, staring at nothing for the moment, is promptly rerouted via c3 and b4 to c5, where it is excellent.

What follows is a long technical phase. 17...a5 18.bxa5 Rxa5 19.Bc3 Ra7 20.Qd2 Qc7 21.a5, and the heavy pieces leave the board down the a-file. Such exchanging phases are not idling: the side with more space loses less in a trade, because the remaining pieces have more squares.

From move 27 the second plan begins: 27.f4 and 29.g4 build the pawn roller, while the bishop on c5 keeps the black knight and the d6 square under control. Objectively the position is only slightly better for White, but it is unpleasant to defend, and against a lower-rated opponent that is the whole point.

Then comes 29...g6, the mistake of the game. Black wants to control f5 and h5, but he weakens f6 and h6 and puts a pawn on a square the white pawn storm can hit directly. 29...Qd8 was the move, eyeing g5 and h4. After 30.f5 the position is suddenly around two pawns better for White: fxg6 and fxe6 are both in the air, and the black king sits in the draught.

The principle that counts here: you do not move the pawns in front of your king to control squares, only when there is no alternative. Every pawn move in front of the king is one more square it can no longer cover.

After 30...exf5 31.Qg5 Re8, the move 32.h3 would have won without any risk at all. What was played is 32.e6, which creates a passed pawn but gives up the c4 bishop's diagonal and the e6 square. Black could have held the balance with 32...fxg4. Instead came 32...f4, and now the position was ripe: 33.Qf6 would have decided immediately, because the queen hits f7 and h8 at the same time.

Colin chose the slower but perfectly sound road with 33.exf7+ Kxf7 34.Qh4, and after 34...h5 35.gxh5 gxh5 36.Qxh5+ Kf8 the diagram position appeared. The move played was 37.Bf3, a good move that wins. Fastest was the line 37.Re6 Qe5 38.dxe5 Kg7 39.Qg5+ Ng6 40.Qxg6+ Kh8 41.Rxe8#, mate with the final move. Black resigned after 37.Bf3, since mate follows within a few moves in every line.

What I tell my students about such an ending: when you are clearly winning, you do not have to find the prettiest mate, but you should know that it is there. In mating positions, look first for the moves that take squares from the king, and only then for the checks.