Analysis
The game in detail
IM Nicolas Brunner (FRA, 2400) vs FM Sigurdur Sigfusson (ISL, 2321) · 1-0
Board 6, ASK Réti vs SG Riehen · King’s Indian Defence, Orthodox Variation, Classical System (E98) · 90 minutes for 40 moves, then 30 minutes, plus 30 seconds per move
By Bogner AI Coach AI Coach . This analysis was written by a Bogner Chess AI coach.
Replay the game on LichessBoard six, ASK Réti against SG Riehen, round eight, all five matches of the round in one hall in Geneva. Across the board sits Nicolas Brunner, international master, 2400, seventy-nine rating points higher, and it is the second round running that an international master with about eighty rating points in hand sits across this board. After seven rounds Riehen are second and Réti fourth, on seven of the eight boards Réti are the lower-rated side, and four and a half board points win the match. Nobody asks board six with Black for more than a half point, and that half point is in the other seven boards’ arithmetic.
It became a loss in twenty-nine moves, and one single recapture decided it, a move Sigfusson spent a minute on.
The storm, and the rook to h6
King’s Indian, classical main line, and after it is only about who gets there first. White runs on the queenside, Black on the kingside, and where both sides are running the number behind the opponent’s name stops mattering. , , , : that is the plan of this variation, Sigfusson plays it without a detour, and costs him 8 minutes 16, with 1:16:39 left after the move. White is a little better, not a whole pawn.
is the first move that costs something, and it is about half a pawn. The rook leaves the f-file to come to the kingside via h6. The other road was , the king on h8, the rook behind its own g-pawn, and the push only after that. Almost nobody at this level plays that, which is why is a decision rather than a reproach.
After comes , and the rook now stands on a square that only one of its own pawns keeps open for it. The diagonal c1 to h6 runs across f4, the black pawn stands on f4, and the white bishop is still on c1. was better: , Black is a pawn worse and everything holds together. A rook behind one of its own pawns is safe exactly as long as that pawn stays there, and for a pawn that is supposed to join the attack, that is not long.
IM Nicolas Brunner (FRA, 2400)
FM Sigurdur Sigfusson (ISL, 2321)
Thirteen minutes for 18...g4
offers the trade on f4, and leaves the offer standing. Brunner spends 17 minutes 08 on the queen move, with 54:48 left after it, and hands a little of his edge back with it: waiting had been working for Black. Sigfusson takes 13 minutes 14 and plays , with 45:12 left after the move. It is the move the plan asks for, and one of the most common choices here; the best move in the position is one almost nobody at this level finds.
It is wrong all the same, and the reason is on the board one move later: removes the wedge, and a white pawn stands on f4 afterwards. Waiting was better, the waiting White had just been struggling with: lets the trade happen when Black wants it, material stays level and White is about a pawn better. was fine too. Objectively the difference is about half a pawn, so hardly a catastrophe.
Practically it is bigger, because from on every move is a capture with a calculation behind it, and the clock had already paid 13 minutes for sending the pawn forward. A pawn that advances gives up a square. The question before was who stands on f4 after , and the answer was White.
IM Nicolas Brunner (FRA, 2400)
FM Sigurdur Sigfusson (ISL, 2321)
19...gxf3, a tactical error
After Black has three captures, two of them take on f4, and one of the three is about a pawn better than the other two: . After 1 minute 04 one of the other two was played, . That is a tactical error, not a question of move order. The pawn goes from g4 to f3 and gives up the square f4, and f4 is exactly the square White needs: attacks the knight on g6 and in the same move opens the diagonal c1 to h6, with the rook at the end of it. Brunner found after 16 minutes 44, with 36:00 left after the move.
was the move, and not as a matter of taste: with a black pawn on f4 there is no f5, and the rook on h6 keeps breathing. After White is a pawn up and stands a good pawn and a half better. That is a game you defend, and most players at this level recapture that way too.
After the rook hangs, and what followed was the practical choice rather than the objective one. takes the bishop and gives up the rook, White wins rook for bishop and stands a good three pawns better; nearly everybody plays it that way. Sigfusson thinks for 12 minutes 46 and plays , which covers h6 and keeps material level but costs about another pawn. Three moves earlier he was a pawn worse, now it is about four. On a capture, what counts is not which pawn you take but which square you vacate.
IM Nicolas Brunner (FRA, 2400)
FM Sigurdur Sigfusson (ISL, 2321)
Three moves that hold, and 22.Bxf4
attacks the knight on g4, and Black finds the answer that holds things together: , after 1 minute 23. The second knight puts itself on the square the diagonal to h6 runs across, defending the rook and offering itself for a trade. There is nothing better here; comes to the same thing, is clearly worse.
Now it was Brunner’s move, and here he did not find the stronger of his two captures. leaves him both bishops and two pawns up. After 3 minutes 03 he played the other one: , and with that both white bishops are off the board against both black knights. White is two pawns up and clearly winning, but after he would have stood about a pawn better four moves later.
That is the one place in the game where Black gets something back, and he got it for three moves: , and are all three the best move, played in 1 minute 23, 1 minute 24 and 3 minutes 53. The rook on h6 is still standing after all of it, and the two bishops that would have collected it are not. When two pieces can capture on the same square, the decision is not the pawn you win but the piece you keep.
IM Nicolas Brunner (FRA, 2400)
FM Sigurdur Sigfusson (ISL, 2321)
The last moves, and 16 minutes for 25...Qe8
arrives after 22 seconds and is one move too early: is the same idea with the king out of the file first, and that is worth about a pawn. After 7 minutes 25 Brunner answers with , which was not the best either and hands a little back.
costs 16 minutes 09, the longest think of the game, and leaves 11:11 on the clock. The move costs about a pawn. was it: takes one of the two white rooks off the board, which is exactly what a player two pawns down wants. After that there is little left to hold against him. , and are not the best moves, but the moves that would have been better are ones almost nobody finds in these positions, and the three-pawn deficit was on the board before them. Sigfusson first went under five minutes on move 28 and never under two, and after he resigned: the clock did not lose this game. Minutes belong to the moves where the game is still open, and here that was move 19 and not move 25.
IM Nicolas Brunner (FRA, 2400)
FM Sigurdur Sigfusson (ISL, 2321)
Verdict
A loss on board six against an international master rated seventy-nine points higher, and with it the team is missing the half point this board was budgeted for: the four and a half board points now have to come from the other seven. One move turned the game, , and it cost 1 minute 04 after the move before it had cost 13 minutes 14. The pawn on f4 was the one holding the diagonal to Black’s own rook on h6 shut, and would have kept it there; the lesson is the same one as in round 7, this time on a capture rather than a queen move, and it runs: what counts is not what you take but which square you vacate. Brunner belongs in the record, because he spent his time where the game was, 17 minutes on move eighteen and 16 on move twenty, and is the move that wins; cost him about a pawn and no more. What Sigfusson keeps are three correct moves in the mess and a game he played to the end without a minute of time trouble. For the team it is a zero on a board where a half point was planned, for him a game lost neither to the clock nor to the seventy-nine rating points, but to a capture that vacated a square.
Moves
1-0
Was this analysis useful?