As noted in yesterday's post on the World Cup, today's tiebreaks featured a lot of top players, and only one of the matches had a clear favorite going in. Sure enough, the matches were all quite difficult and one even reached the Armageddon stage.
Let's start with a recap of the G/25' + 10" action. Veselin Topalov was the one clear favorite alluded to above, and in his first game with Lu Shanglei he won easily when his opponent blundered an exchange in a position that was already pretty lousy. But then it got more interesting. For the vast majority of the game Topalov was in no danger. He'd have a winning position, then he'd let his opponent slip out and be okay, then he'd be winning again, then even, then winning - this happened quite a lot. As the game gradually worked its way into an ending, however, Topalov slipped up in a serious way and allowed Lu to achieve rook and knight vs. bishop and knight, with no pawns. It was still a theoretical draw, but Topalov was rattled and was getting into real danger of losing the game. Just in time, though, he came up with a nasty trick. He spent a precious minute on move 87 to come up with 87...Bf2, setting a little trap that Lu fell right into. After 88.Rh2 Kf3 89.Kxd3 it looked for a moment like White was winning, but 89...Bg3! was the point. White could only try rook vs. bishop for a few moves, but with Topalov's king near the correct corner (e.g. one opposite the color of his bishop) there was no danger, and the match was soon over.
Peter Svidler overcame Teimour Radjabov in a "pick 'em" contest between two essentially equally strong players. Radjabov appeared to have a serious advantage in his white game, but when Svidler reached the exchange-down ending with all the pawns on the a- through e-files it was a clear draw. Radjabov continued the game for a long time, but without any real chances to win. The second game was a bit funny, in that Svidler had White against the Gruenfeld, and he played a line Radjabov himself had used with the white pieces. Apparently Svidler dealt with the role reversal better than his opponent, and soon Radjabov had no compensation for the sacrificed pawn. In general Svidler won a clean and smooth game, but there was one serious hiccup. His 26.Ne6 is objectively a '??' move, as it took him from a winning or near-winning game to a lost one. Radjabov had to find one key tactic, however, and he didn't. He took with the bishop and lost without a fight, but 26...fxe6 27.dxe6 Rxc4! would have resulted in an easy win, as Black's bishops would have overmatched the rook.
Wesley So also made it through to the 4th round in the G/25s against Le Quang Liem, though not by traditional means. In the first game, he got nothing with White in a Berlin ending and was the weaker half of the draw. In the second game he was a little worse, as Le seemed to have nagging pressure in a position where White couldn't possibly lose. No way at all...or so it seemed. In fact, it was a sort of trap, and Le fell headlong into it with the natural 30.Kf4?. So's 30...h5 won a pawn, or at least that's all it would have done if White had understood what was happening. Instead, he played 31.Rc2??, and after 31...h4 he could have resigned. The threat is 32...g5#, and if 32.g5 Rf5# is another mate. White sacrificed his knight to avoid the mate, but the resulting endgame was hopeless and he resigned a few moves later.
The last match to be decided in the 25-minute games was between Vladimir Kramnik and Dmitry Andreikin. With the one exception of the 2013 World Cup final, Andreikin has been besting Kramnik on a regular basis, and he did so once again in today's games. As usual in rapid games, there were adventures. The first game was relatively equal through Kramnik's 30.Nd4, and then Andreikin uncorked the horrific 30...Rd6, walking into the obvious one-move fork 31.Nc8. Andreikin could have resigned there, but probably from a combination of disgust and both sides being fairly low on time he played on. Nevertheless, down the exchange and, after several more moves, down a pawn on top of that, he battled on, and somehow Kramnik couldn't figure out how to win. Kramnik has had many high-profile failures with his technique over the years, going back to at least his Candidates' match with Boris Gelfand back in 1994, but this might be the biggest failure of his career in terms of the size of his advantage and the degree to which the win should have been completely routine. Full credit to Andreikin though; he hung in there and kept causing problems. After Kramnik's 52.Rxe4+? (instead of first playing 52.c6 to pull Black's rook off the second rank), the game could no longer be won.
It often happens that after such a failure, it's too hard to recover, and indeed Kramnik did not manage to do so. Kramnik's pawn sac in the opening of the second game was either a sort of "what the heck?" move that showed that the wheels had come off, or else he didn't manage to remember his prep. Whatever the case, he was soon down a pawn with a lousy position, and with fine endgame technique Andreikin managed to win an opposite-colored bishop ending to advance to the fourth round.
Now to the longer matches. With White Evgeny Tomashevsky started the playoff with Maxime Vachier-Lagrave (MVL) with excellent winning chances in the first 25-minute game, but he twice let his opponent off the hook. (46.Bf2 was a good opportunity to keep playing for the win.) MVL had a token advantage in the second game, but there was never any serious chance to convert with the opposite-colored bishops. MVL broke through in the first 10' + 10" game, winning with White. They played the same opening as in the G/25, and if anything Tomashevsky enjoyed an even better position in the middlegame the second time around. The key moment came on move 30, when 30...Qd7 would have given him a good game. Instead, 30...Rf8? left White with the advantage, and he was soon able to convert it into a full point. The second game was balanced for a long time, but Tomashevsky's need to mix things up eventually led him to overpress, and MVL won the second 10-minute game as well, winning the match by an overall score of 4-2.
The match between Michael Adams and Leinier Dominguez went even deeper, to the 5' + 3" games. Neither player missed any big opportunities in the 25-minute games, but Adams apparently missed a fairly simple tactic in the first 10' + 10" game that would have netted an exchange and most likely the game: 23.Bxf5 Bxf5 24.Nd7. He missed an even bigger chance several moves later, though it was also more subtle: 28.Bxf5 Bxf5 29.Qxf5. That part is obvious, but so is Black's most natural rejoinder, 29...Qxg3. The subtle part is to realize that 30.Qf6 simply ends the game: 30...Bxe5 31.dxe5 Q-anywhere safe and then 32.Bh6 followed by 33.Qg7#. (I'm sure that if Adams had a second, he went out of his way to make sure that Adams never found out about this until after the match.) The second 10-minute game was another fairly innocuous draw, and then it was on to the 5-minute games.
In the first one, something incredible happened. Adams was gradually outplayed, despite having the white pieces, and eventually lost the exchange (or blundered it, if one goes back a couple of moves before the fork that won it) for no compensation whatsoever. But somehow he managed to create a fortress, or at least a reasonable enough facsimile thereof, and Dominguez couldn't find a way through. Eventually Dominguez opened the queenside, but somehow that only helped Adams to obtain dangerous counterplay. The game was already unclear by the time Dominguez blundered with 84...Qb4??, and that allowed Adams to regain his exchange - with two extra pawns. (Blundering an exchange...where have we heard that before? It seems to be the theme of the day.) Adams won in the sequel as well, and qualified for the next round with a 5-3 victory.
Finally, there was the most exciting match of the day, contested by the runners-up in last year's world blitz championship, Hikaru Nakamura and Ian Nepomniachtchi. (In fact Nakamura is still #1 in the world in rapid, slightly ahead of Magnus Carlsen, while "Nepo" is #10; in blitz Nakamura is #2 and Nepomniachtchi is #4.) In the first rapid game, Nakamura had some chances with White, but Nepo defended well and managed to hold. The same was true, with colors and players reversed, in the second game. Nepomniachtchi did have one interesting possibility there, but it probably wouldn't have won: 38.e5+ Kxc6 39.b5+ Kb7 40.bxa6+ Ka7 41.Kd4 Bxa6 42.Bh5 Bc8 43.Bxf7 Kb7, and now White doesn't have anything that's obviously better than 44.f5 exf5 45.e6 Kc7 46.e7 Bd7 47.e8Q Bxe8 48.Bxe8. If White's remaining pawn were on any other file, he might be winning, but here he's left with a wrong-colored bishop and rook-pawn combination. Black puts his king on h8, waves his pawns goodbye and draws in his sleep.
After this, all the games were decisive. Nakamura won the first G/10 after Nepomniachtchi fell for an elementary trap in time trouble; the problem, of course, is that in time trouble even elementary traps can be deadly. That was an see-saw struggle, and so was the second game. It looked like Nakamura was going to win the match, as he had taken the upper hand in a very complicated game, but then he made a huge error that at first seemed to be very strong. His 40th move, 40...e4, wins against every White move but one, and that's what Nepo played: 41.Rd2! Black is completely busted after this, and soon the match score was leveled.
The first five-minute game saw further adventures. Nakamura again eventually managed to get the advantage with Black when White blundered - you guessed it - the exchange - but it wasn't such a big advantage this time around. In fact, winning the exchange did Nakamura a disservice. Rather than playing it safe and going for a draw a few moves later when it was objectively clear that Nepomniachtchi wasn't in any danger, Nakamura overpressed and was soon losing the game.
A lesser player may have folded after that, but if Nakamura is anything he's mentally tough, especially as a blitz player. Nakamura won the next game, and while it wasn't a perfect game it was probably the cleanest victory of the entire match. As a result the players moved on to the Armageddon game, with Nakamura taking Black, four minutes and draw odds against Nepomniachtchi's five minutes. Nepo had his chances early, but Nakamura's kingside buildup proved more effective, and even before White fell into a mate at the end Black was clearly winning and in no danger whatsoever of losing the game.
Tomorrow, round 4 begins, with the following pairings (given in bracket order):
All three United States players are still in it, and if they keep winning they won't face off in the quarter-finals either - good news for us. There are four Russians still in it, and the home country (Azerbaijan) still has one representative in the fight. Surprisingly, only two Chinese players are left, and they're facing each other next: China's current #1 against the player of their future - which might turn out to be now. The tournament has been very hard on the veterans so far, but three of them are still in it. Unfortunately for them, two of them are facing off against each other and the third is facing Nakamura, but who knows what will happen? Adams is a tough and tricky guy, but I think that unless Adams wins in classical chess Nakamura will manage to catch him with tactics in the quicker games. Another problem for the older players is that there are no days off (except by winning in the classical games or by taking de facto days off with quick draws in the classical games) until after the quarter-finals. Then, at long last, the final four will get to enjoy ONE day off.