World Cup 2015: Round 7 (Finals), Day 5: Karjakin Wins a Long and Crazy Tiebreaker
The 2015 World Cup came to a madcap conclusion after six crazy tiebreak games, all of them decisive, and after a series of topsy-turvy games the last man standing was Sergey Karjakin. Both he and Peter Svidler have qualified for next year's Candidates' tournament, but it was Karjakin who secured the title of World Cup Champion and the extra $40,000 that went with it. He won $120,000 to Svidler's $80,000, and both will be able to afford some quality help as they try next year to qualify for a match with Magnus Carlsen.
The tiebreak match was every bit as crazy as the classical match (or rather, the classical portion of the match). Both players won must-win games, and knowing what happened in the preceding game or even who came out of the opening better off told one almost nothing about what would happen by the game's end. Let's give a quick recap of the games.
In game 1 (game 5 overall), the first of the 25' + 10" contests, Karjakin had White and played an unusual line of the English. He may have obtained a slight edge at first, but Svidler outplayed him and managed to achieve all his strategic aims. Perhaps he was never quite winning, but he was much better for a very long time. He was never in danger until he missed 42.Ng4, and even then things weren't so bad for a while. Finally, though, they reached an opposite-colored bishop ending where Karjakin didn't seem to have any real winning chances, but he managed to make progress. He was a little careless about it and gave Svidler three chances to find a blow that would have put Karjakin on his heels trying to find the draw, but finally White put a stop to the idea. After 79 moves the last critical moment of the game appeared. Karjakin had made as much progress as he possibly could "for free", while Svidler was at his last line of defense. Unfortunately for him, Karjakin had, and found, 80.d5!!, and that won the game.
Svidler had lost three games in a row by this point, one he was winning outright (game 3) and one (this one, game 5 overall) where he was much better for a very long time. Now it was his turn to have to win to stay in the match, and despite all the discouragement he must have felt by this point he came through in spades, winning a remarkably clean game with White in a King's Indian Attack. Somehow the Svidler of the first two and 9/10 games had returned, and it was match on once again.
Games 3 and 4 of the tiebreak (7 & 8 overall) were with the 10' + 10" time control, and this time Karjakin faltered with White. His attempt to avoid an inferior sort of Modern Benoni led to something worse, and things just snowballed from there. Karjakin was already losing when he played 23.Rxc4?, but that cost him a piece to an elementary tactic. The game went on a while longer, probably so Karjakin could get mentally prepared for the next game, but there was never any doubt as to the result over the course of the remaining moves.
So once again Svidler had White in a game where he only needed to draw, and once again...he failed to score. He chose a differet anti-Sicilian line than he did in game 3, but here too it wound up in a sort of Maroczy Bind. Karjakin apparently surprised him somewhere, because Svidler played hesitatingly and then overreacted and overreached on the queenside. After 15 moves Svidler was losing a pawn, after 23 moves he was losing a second pawn, and after 27 moves he resigned. A disaster for Svidler, but good nerves by Karjakin, who had saved his third "match point".
On to the blitz: 5' + 3". Yet again Karjakin started with the white pieces, and this time there was something new. Karjakin played 1.e4 for a change, and Svidler headed for his beloved Marshall. Karjakin went all the way into the rabbit hole, but even though they went into a main line he was somehow unprepared and played a terrible novelty on move 18, Bc2. Svidler had plenty of time, having gained almost 40 seconds on the clock by this point, and he used two or three minutes trying to figure out what to do against this new move. It's clear from his facial expressions that he figured it out...a few seconds after making the wrong move. Black could have played 18...Nxc3, with a decisive advantage. A less tired Svidler would have spotted this (for that matter, a less tired Karjakin probably would have seen it as well before playing 18.Bc2), but instead he played 18...b4, which was a bad move in its own right in addition to missing a huge opportunity. A few moves later Karjakin could have been winning, but even with his imprecise 24th move he maintained a significant advantage.
Unfortunately for Svidler, his horrors had not yet come to an end. Bit by bit he fought back, only to make a serious error on move 28. Karjakin could have replied to 28...Bh5 with 29.g4, with a winning advantage, but instead he played 29.Rb1??, allowing several winning rejoinders. Svidler's wasn't the best, but it was good enough for a winning advantage. Once again it was time for Karjakin to demonstrate his resilience, and while Svidler maintained his advantage over the next dozen or so moves he didn't make much progress. He did have an advantage on the clock, but having more time and more material won't rescue you when you blunder a rook, as he did. He simply left his rook en prise on b8, Karjakin took it, and Svidler's anguish was palpable. You will rarely see a player more distraught than Svidler was, but there was still another game to play, one more chance for a comeback.
With White Svidler managed to achieve a serious, possibly winning advantage. This time around there weren't any blunders that gave the game away; it was just the slow but steady drip of inaccuracies that allowed Karjakin to equalize, and at the end even win when Svidler was forced to gamble to avoid a draw. Svidler found a very nice trick at the end of the game that could have won, but Karjakin found the right response and Svidler had to resign. A very bitter end for Svidler, who was winning or nearly winning in games 3, 5, 9 and 10 and lost them all.
So Karjakin won the tiebreak 4-2 and the match by the overall score of 6-4. (The games are here, with my comments.) Incredibly, both players are headed to Berlin at the end of the week for the World Rapid & Blitz Championships. They might be so tired by now that I would stand a decent chance against them, but hopefully they will recover well enough to play near enough to full strength to avoid a catastrophe. One final note about the event. At the end of the press conference Sergey Karjakin mentioned that he had been given some advice by Sarkhan Gashimov, the brother of the late super-GM Vugar Gashimov and a talented chess player in his own right. He also recounted his friendship with Vugar and that the dying Vugar had told him that from now on he, Karjakin, had to play for the both of them. It was a moving moment and a nice gesture to the Azeris, as Gashimov was one of their own. We shall see next March whether he, or Svidler, can take a further step towards the world championship when the Candidates' tournament takes place.
Reader Comments (5)
It was physically exhausting and emotionally draining - just to watch. Svidler already won once and Karjakin is about to become a father so I guess one could rationalize he "deserved" the victory.
[DM: Right - and you can throw in the Gashimov tribute as well.]
Tuesday morning quarterbacking: Svidler's poor clock-handling was a major factor at the end: in winning positions he had too little time to find the decisive blow and when Karjakin was in severe zeitnot played too fast to try to hustle him.
[DM: I don't think so. He had plenty of time, for instance, when he failed to win game 3 and when he blundered the rook in the next to last game of the tiebreak. Also in that game he had tons of time to spot the winning ...Nxc3 after Karjakin's blunder of a novelty. This was about nerves and exhaustion rather than zeitnot.]
I wonder: is this what the people want?
Dennis, thank you for the fantastic coverage of the event, much appreciated.
One question about the first game: is the bishop ending after 80.d5!! really winning for White? Some commentators thought it's still drawn after 82...Kd5 (GM Landa suggested 82...Kd5 83.Ke7 Kc4 84.e6 Kxc3 85.Kf6 b4 86.e7 Ba4 87.Kxg6 Be8 88.Kg7 b3 89.Kf8 Ba4 90.Be5 Kd3 as the drawing line). Engines are often confused in these opposite color bishops positions.
[DM: Andrey, thanks for the kind words and the good question. I agree that computers can be confused, so it's important to understand what evaluation numbers are meaningful and what ones aren't. I actually spent a bit of time on that ending to make sure it really was winning after 80.d5, and although I decided not to post my work in that case (with six games to cover I wanted to keep things manageable!) I'm at least reasonably sure that the ending is won for White.
After 82...Kd5 (instead of Svidler's 82...Bc4) 83.Ke7 Kc4 White should play 84.Kf6 (or 84.Kf7) rather than Landa's 84.e6, and this seems to win:
(1) 84...Kd5 85.e6 Kc6 86.Bg5 Bd5 87.e7 Kd7 88.Kxg6 is essentially the same win as at the end of the game.
(2) 84...Kxc3 85.Kxg6 b4 86.Kxh5 Bf7+ 87.Kg5 Kc2 (If 87...b3 88.Bc1 Kc2 89.Ba3 wins easily) 88.h5 b3 89.e6 b2 (89...Bxe6 90.Be5 wins) 90.exf7 b1Q 91.f8Q+. Black has also managed to queen, but winning for White is elementary.
It's possible that I missed something here, or that Black has some other try, but at this point it seems to me that White is winning after 80.d5, and maybe as far back as 69...Be6? rather than 69...g5.]
Thanks for the coverage, Dennis. I have a question on tiebreak game 2: Can white make progress if black stays passive rather than "activating" his rook with 37.-Ra3, to decisively activate the _white_ rook? I think the only way would be to push his kingside pawns, with queens on the board this looks double-edged (but Svidler probably had to do it eventually in a must-win game).
[DM: If you mean decisive progress, I'm not sure. You're right that Black should have stayed put and hoped for the best, as difficult as that is to do.]
In addition to blunders, the entire match had interesting moments: when to switch betwen passive and active defense (e.g. also classical game 4 where Svidler remained passive which was wrong), when to open the position (classical game 1: decisive for white, classical game 2: OK for black, tiebreak game 3: horrible for white). I also have trouble "believing" the e3 Benoni, but it had been played before at the highest level with rather good results.
[DM: Not all e3 Benonis are created equal. When White's dark-squared bishop is outside the e-pawn it's not so bad, but it's less impressive when it's behind the pawn. But even then it has its drop of poison. I remember this game (I was there at the time), which pretty much put an end to Kramnik's chances of keeping his world champion's crown: http://www.chessgames.com/perl/chessgame?gid=1473793]