GM Alexandra Kosteniuk (SUI, 2495) vs IM Marina Brunello (ITA, 2308) · ½-½
Board 1, Switzerland vs Italy, women’s event · Sicilian Defence, Moscow Variation (B51)
By Professor Zugzwang AI Coach . This analysis was written by a Bogner Chess AI coach.
Round 6 of the women’s event, table 13, Switzerland against Italy, both teams arriving on eight match points: on board one GM Alexandra Kosteniuk had White against IM Marina Brunello, rated a hundred and eighty-seven points lower. It became a Moscow Variation, and as early as move four the bishop stepped aside to a4 instead of trading on d7.
Fifty-four moves later the game was drawn. In between lies a bishop sacrifice on h6 on move 22, a won position after move 29, and on move 30 one step of the king that handed all of it back in a single moment.
When the centre closes: 15.f5
After White chooses between the trade, the retreat to e2 and the step aside to a4. In four games out of ten White simply castles here. Kosteniuk keeps the bishop on the a4-e8 diagonal and builds control of d5 with : the freeing push d6-d5 becomes work for Black from here on, and in this game it did not arrive until move 40.
planted a black pawn on c6, where it blocks Black’s own a8-h1 diagonal, and gave Black the half-open b-file in return. asked the question, answered it by bolting the centre shut, and that made the strongest move on the board. The pawn takes e6 away from the c8 bishop and g7-g6 away from the king, and from that moment until the last move of the game it stood on f5.
The bishop on c8 paid for it. It went and and spent its time looking at its own pawn on c6. The rule behind this is old and it holds: with a closed centre you play on the wing your own pawn chain points at. After f5 White was pointing at the kingside, and the rook set off, , , . Straight to h3 with was a shade more accurate, but the detour via g3 cost little.
IM Marina Brunello (ITA, 2308)
GM Alexandra Kosteniuk (SUI, 2495)
Two attackers, one defender: 22.Bxh6
Count the attackers on h6: the bishop on e3 and the rook on h3. Count the defenders: the g7 pawn and nobody else. Two against one, and the black position still held while the king stood on h8 and the bishop on e7. The reason is the line : the f8-g7-h6 diagonal is open then, and the bishop covers the very square the capture happened on. does the same job with the knight.
The game saw , and a king does not defend h6. That turns a pleasant edge of about a pawn and a half into an attack worth a good three pawns more, and stops being a question of nerve and becomes a question of counting: . White has given a bishop for two pawns and has a black king standing in the open in the middle of the board.
The quiet would have kept a large advantage too, but the sacrifice is worth roughly two pawns more than that. This is exactly the kind of calculation you can train, and it is simpler than it looks: count the attackers and the defenders of the square first, then ask what the captured pawn was doing for the king, and only last of all look at what the piece is worth. Calculate in that order and finds itself.
IM Marina Brunello (ITA, 2308)
GM Alexandra Kosteniuk (SUI, 2495)
The useful move instead of the forcing one: 30.Nd5
After everything is in place. The knight on f6 is attacked by the queen on g5 and defended by nobody, and attacks it a second time. Capturing it is out of the question: wins the knight back, and since c6 has been vacated on the way, the bishop on a4 now looks at the rook on d7.
Black has nothing better. hands over the queen for rook and knight, tears open the last file and leaves the black king in the middle. That is not a position with an advantage any more, that is a won game.
The game saw , and the move is no whim. On h1 the black queen had checks on e1, f1 and g1; on h2 she had none. Prophylaxis is a virtue, and here it cost the whole game: put the position back in balance. The fine detail is that the same king step appears in the winning line, one move later and with a gain of time. In positions like this, look at the forcing move first and at the useful one afterwards, never the other way round.
IM Marina Brunello (ITA, 2308)
GM Alexandra Kosteniuk (SUI, 2495)
The passed pawn runs too early: 35.h7
On move 31 White had let the position slip once already, and Black gave it straight back. would have turned the game around and left Black about two pawns better; the game saw , a pawn taken at the cost of her own king. After that the h-pawn marched to h6, the black queen shuttled between f2 and e1, and White was clearly winning again.
looks like the crowning move and is the mistake. trades the queens with check, and without queens a pawn on h7 is only a pawn: picked it up and material was level again. instead of was the last chance and would have kept about two pawns.
The move was . The rook attacks the queen on f2, so she has to leave, and it covers f4 a second time, which takes the trade with check off the board. Only then does the pawn run, for instance , and the pawn queens with check. A passed pawn does not become strong by running, it becomes strong when the opponent has no counter left. Take the checks away first, then push.
IM Marina Brunello (ITA, 2308)
GM Alexandra Kosteniuk (SUI, 2495)
Verdict
The end was a bishop ending with both bishops on light squares, two white pawns against one, the white pawns doubled on the f-file and the black king in front of them: , and the hands went out across the board. For Alexandra Kosteniuk it is the third half point in five games of this Olympiad and the most annoying one, because the game tells the story of three decisions. The bishop sacrifice on h6 was completely correct and deserved a full point. The king step on move 30 was useful and gave a won position back. The passed pawn on move 35 ran one move too early. Marina Brunello helped generously with , and , and at the board the win does not go to whoever had more chances but to whoever finds the forcing move at the right moment. On board one with White against an IM half a point is still no bad result, and in a team match it counts exactly as much as any other.
Moves
½-½
Was this analysis useful?