26 May 2015
Grandmaster Vasily Yemelin analyses the games of the second round of FIDE Grand Prix in Khanty-Mansiysk.
The Grand Prix's second round apparently kept its online audience glued to their screens. Most of the games ended with a score that could hardly be expected from the logic of the encounters. Well, the event is becoming ever more interesting.
Svidler – Caruana
1.d4 Nf6 2.c4 e6 3.Nf3 d5 4.Nc3 Be7 5.Bg5 h6 6.Bh4 0–0 7.e3 Ne4 8.Bxe7 Qxe7 9.cxd5 Nxc3 10.bxc3 exd5 11.c4 Be6 12.Rc1 Nd7 13.Qb3 c6 14.Be2 Nb6 15.cxd5 Bxd5 16.Qb2 Na4 17.Qd2 b5
Up to now, the players had been following the rapid game Grischuk – Radjabov, Beijing 2014, in which Black continued 17...Qa3, and White provided a poor response. It remains to be seen what Caruana meant by avoiding this line. For now, I can make an assumption that after 18.0–0 (18. Rc2 Be4 is pointless) 18...Qxa2 19.Rc2 Qb3 20.Ra1 White can expect to gain a slight edge because Black pieces become entangled. Anyway, White regains the pawn soon: 18...Rfe8 21.Bd1 Bxf3 22.Bxf3 Nb6 23.Rca2 Nc4 24.Qc1 Nd6 25.Rxa7 Rxa7 26.Rxa7.
18.Ne5 Rac8 19.0–0 b4 20.Ba6 Rc7 21.Nd3 Rb8 22.Nc5 Nxc5 23.Rxc5 Qg5 24.g3 Qg4 25.Bc4 Bf3 26.Rc1 Rcb7 27.Qd3 g6?!
By playing precisely, White was able to exploit the strengths of his position, while Black's threats on white squares proved to be insignificant. On top of that, Black's last move turns out to be highly inaccurate, which enables White to gain a decisive advantage: 28.Re5! White's plan is to play against the bishop on f3, which doesn't have enough squares. A simple e4 with the following h3 threatens, and there is no convincing defense. An attempt to chase away the rook does not help in the least: 28...Kg7 29.e4 f6 30.Re6 f5 31.exf5 Qxf5 32.Qe3 with the threat Bd3.
Unfortunately, this beautiful solution didn't occur as White opted for a calmer continuation.
28.Bb3 Be4 29. Qd1
White could have forced a trade-off of the queens with 29. Bd1, but the move in the game, offering an obviously better version of such a trade-off, is stronger. By abstaining from trading the queens, Black only worsens his position because White's center is shifting.
29...Bf3 30.h3! Bxd1 (30...Qe4 31.Bc2) 31.hxg4 Bxg4 32.Rxc6 would have resulted in a very bad endgame.
30. f3 Bd5 31.e4 Bxb3 32.axb3 Rb6 33.Qd3 Rd8
Black's position after the obvious 34.R1c4! is both bad and ugly. A blunder followed, however.
34.d5? cxd5 35.Rxd5 and, without waiting for the "surprise" move 35...Rd6, which, as I think, was no longer a surprise for Svidler as soon as he took the d4 pawn in his hand, White offered a draw that was accepted.
Gelfand – Nakamura
1. d4 Nf6 2.c4 g6 3.Nc3 Bg7 4.e4 d6 5.Nf3 0–0 6.Be2 e5 7.0–0 Nc6 8.d5 Ne7 9.b4 Nh5 10.g3 f5 11.Ng5 Nf6 12.Bf3 fxe4
The line is known for wins that Kramnik, the main expert of this variation, scored against Grischuk and Giri after 12...c6. Along with the move in the game, the immediate 12...h6 also occurred in the past.
An interesting subtlety compared to the previously played 13...Nf5 14.Bg2 (nothing can be gained by 14.Ne6 Bxe6 15.dxe6 Nd4 16.Nxf6+ Qxf6 17.Bxb7 Rab8) 14...Nxe4 15.Nxe4 a5 16.b5 b6. In this type of positions, White has some initiative thanks to the only available strengthening plan f2-f4, for example: 17.Bb2 Bd7 18.Qd2 Qe7 19.Rae1 Rae8 20.f4 Kh8 21.Ng5 h6 22.Nh3, Ipatov – Mamedov, Antalya 2013, and White has a slightly more agreeable position.
The subtlety after 14.Nxf6+ Bxf6 15.Ne4 Bg7 apparently needs to be explained. I believe that the idea is to prevent the knight from immediately approaching d4, as well as the move g6-g5, for example: 17.Bg2 a5 18.b5 b6 19.Bb2 g5. But the standard intrusion on e6 is more straightforward, of course.
14...Bxe6 15.dxe6 Qc8
The continuation 15...Nxe4 16.Bxe4 c6 17.b5 fully illustrates White's ideas in the following line: after 17...Rf6 18.bxc6 bxc6 19.Rb1 Rb8 20.Ba3 Rxb1 21.Qxb1 Qc7 22.Qb3 Rxe6 23. Rb1 the extra pawn doesn't make Black's life easier: his center is weak, and White has a strong initiative.
White probably decided to keep the queen on d1 to prevent c7-c6 for a while, but now Black finds a good opportunity to win the pawn e6 without any real positional cost. There was also a pretty good option 16.Qb3 Kh7 (a similar position occurs after 16...Nxe4 17.Bxe4 c6 18.Rd1 Rd8 19.b5 Qxe6 20.Ba3) 17.b5 Nxe4 18.Bxe4 c6 – compared to 16...Ne4 Black doesn't have to put his rook on d8, but this barely affects the evaluation of the position since White has a good compensation for the pawn.
Black could have chosen a primitive 16...Nxe4 17.Bxe4 Qxe6 18.Bxb7 (18.Qb3!?) 18...Rab8 19.Bd5 Nxd5 20.cxd5 Qf7 21.Qd2 Rb5 22.Rfd1 a5 23.bxa5 Ra8, with probable trade-offs and a draw ahead, even though White would have maintained some pressure. But of course, Nakamura rejects such a development.
I think Gelfand was too indifferent to the fate of the desperado pawn on e6. The approximate continuation 17.Bg4 Nf5 18.Qd5 (nothing can be gained from 18.e7 Rf7) 18...c6 19.Qd2 Qxe6 20.Bxh6 led to a complicated position with material equality and a slight initiative for White.
17...Qxe6 18.Qb3 Kh7 19.b5 Nf5 20.Rad1 a6 21.Nc3
After 21.a4 White seemed to retain sufficient compensation. But now the game converts into an endgame where White has to struggle for a draw.
21...axb5 22.cxb5 Qxb3 23.axb3 Ra3 24.Rb1 Nxe3
A simple but passive 24...b6 was possible, but Nakamura correctly chooses a forced line allowing him to open up the position and activate the bishop and, with some luck, the knight as well.
25.fxe3 Rxf1+ 26.Kxf1 e4!
One thing is clear: if it works, play it!
27.Nxe4 Ra5 28.b6 c6 29.Rd1 d5 30.b4 Rb5 31.Nc5
This mistake has apparently cost Nakamura serious chances for a win. After 31...Nd6 with the idea 32.e4 (32.Rb1 Rxb6) 32...Nc4! 33.Rd3 d4 or a simple capture of another pawn 31...Rxb6 32.e4 d4 33.Rb1 Nc7 34.Ke2 Na6 Black has a pawn up and a clear advantage. White has no chance of catching the b6 rook in either case.
Black could have captured on e4, but after 33.Rd7 Rb6 34.Rb7 Rb7 35.Nb7 White soon wins the pawn e4 back, and gets the same minor piece endgame as in the game.
It's amusing how White repays his debt for Black's 26th move. The same comment is relevant here, even more so since after 33.Nxb7 Nf6! with the ideas Ng4 and Nd7 White would have had to face some difficult problems. But now the position simplifies, and it seems impossible to convert the extra pawn into a win because of the opposite-colored bishops.
33...Rxb6 34.Rxd4 Rb1+ 35.Ke2 Rb2+ 36.Kf1 Bxe5 37.Re4
37.Rd7+ would have been an easier way to a draw.
37...Bd6 38.Nxb7 Rb1+ 39.Re1 Rxe1+ 40.Kxe1 Bb4+ 41.Ke2 c5 42.Kd3 Nf6 43.Bf3 Nd7 44.Be2 Be1 45.Nd6 Bf2 46.Ke4 Bg1 47.Kf3 Ne5+ 48.Kg2 Bd4 49.Nc4 g5 50.Nxe5 Bxe5 Draw.
Vachier-Lagrave – Jobava
1. e4 c6 2.d4 d5 3.e5 Bf5 4.Nf3 e6 5.Be2 f6
This is impressive: Black uses a novelty on the 5th move. The strongest player who ever made this move before was rated about 1700.
6.0–0 fxe5 7.dxe5
A more peaceful and natural 7.Nxe5 Nd7 8.Bd3 Qf6 9.Bf4 Bxd3 10.Nxd3 Qxd4 11.Re1, with a minor advantage for White, would have looked fine as well. Vachier-Lagrave opts for a more aggressive and obliging plan involving an attack of the pawn e6 and the bishop f5.
7...Bc5 8.Nbd2 Ne7 9.Nb3 Bb6 10.Nfd4
In such positions both humans and computers have the urge to play 10.a4, although the purpose of this move is not very clear.
The computer favors the line 10...c5 11.Nxf5 Nxf5 12.c4 Nc6 13.f4, but there are great doubts whether the bishop on b6 can be made active.
White certainly had to check the line 11.g4 c5 12.gxf5 cxd4 13.fxe6 Nbc6: Black has a good play since White can't defend the pawn e5. For example, 14.a4 a5 15.Kh1 Nxe5 16.Nxd4 Qd6, etc. with an unclear position.
11...Nd7 12.g4 Bxd4 13.Nxd4 c5 14.Nb5!
It's difficult to choose among the options, but White makes the right decision. A simple 14.Nxf5 exf5 15.f4 seems tempting, but Black manages to place his knights well, the d4 square being a very attractive target: 15...Nb6 16.b3 dxc4 17.Be3 (or 17.bxc4 fxg4 18.Bxg4 Nf5) 17...Qc7 18.Qd6 Rac8 19.Rfd1 fxg4 20.Bxg4 Nf5 with unclear play.
Black also survives in the heated battle after 14.gxf5 cxd4 15.fxe6 (15.f6 Nf5 is bad) 15...Nxe5 16.Qxd4 Nxc4 (16...N5c6) 17.Bg5 (after 17.Bxc4 dxc4 18.Qxc4 Rf6 White can't retain the pawn e6; 17.b3 Nc6 18.Qd1 Nb6 19.f4 Qd6 20.Bg4 d4 leads to a double-edged position) 17...Nc6.
After 18.Qg4 Qd6 19.Rad1 N4e5 20.Qg2 Qxe6 21.Rxd5 (21.Qxd5 Qxd5 22.Rxd5 Nf3) 21...Rf5 22.Be3 Raf8 the position is complicated, and the two bishops do not give White any clear advantage due to the weakness of his king.
After 18.Qh4 Qd6 19.Bd3 (19.Bxc4 dxc4 20.e7 Rf5 21.Rae1 Re8 22.Qxc4+ Qd5 23.Qxd5+ Rxd5 equalizes) 19...h6 20.Bxh6 N4e5 21.Kh1 Nxd3 22.Bxg7 Qxe6 23.Bxf8 Kxf8 White can't involve its rook in the attack and has to settle for a perpetual check: 24.Qh8+ Qg8 25.Qf6.
14...a6!? looks appealing, but after 15.Nc3 (Black is perfectly fine after 15.gxf5 axb5 16.fxe6 Nxe5 17.cxd5 Nxd5 18.f4 Nc6 19.Bxb5 Nd4 and stands even better after 15.Nd6 Nxe5 16.Nxb7 Qb6) 15...Bg6 16.cxd5 exd5 17.Bg5! Nxe5 18.f4 (if 18.Nxd5 Qxd5 19.Bxe7, there is an exchange sacrifice 19...Bd3! 20.Bxf8 Rxf8 with sufficient compensation) 18...N5c6 19.f5 Be8 20.Bf3 Bf7 (it's probably better to give the pawn back – 20...Qd7, but then White has a clear advantage) 21.Qd2 White is able to develop a dangerous initiative.
During the next few moves, White will have to deal with the pressure in the center. It is clear that White is obliged to assess the consequences of complications very carefully because his king is weak. 15.cxd5!? is strong now because 15...Nxe5? is bad due to 16.f4.
Nor is Black out of the woods after 15...exd5 16.f4 Qb6 17.Nd6 c4+ 18.Kh1 Qc5. Now there is a good option 19.Re1! with the idea Bf1 and Be3, but the more primitive 19.Rf3 is also possible (with the threat Be3) 19...Nc6 20.Be3 d4, although here there is a strong pawn thrust 21.b4!.
After 15...Nxd5 16.f4 N7b6 17.Nd6 Nc8 (17...Qe7 18.a4 Rad8 19.Qb3 is worse) 18.f5 Nxd6 19.fxg6 Rxf1+ (19...Ne4 20.Rxf8+ Qxf8 21.Bf3 loses because the knight is caught) 20.Qxf1 Ne4 21.Qf7+ Kh8 22.Bh6 Qg8 Black seems to hold the position by the skin of his teeth, but White is still better.
15.f4 Nb6 16.b3
16.cxd5 Nexd5 17.Nd6 would have led to a position considered in our previous comment.
16...Be8 17.Nd6 Bc6 18.Be3?!
This seems to be a serious positional oversight. White seeks more certainty in the position, but changes it in Black's favor, and the chances equalize. A tempo move 18.Ba3 would have been strong, although the bishop is obviously not in its proper place. After 18...Nd7 19.Qd2 Nc8, however, White's idea 20.cxd5 Bxd5 21.Nb5 leads to a better position. White should have prevented the center from getting blocked.
18...d4 19.Bd2 Nbc8 20.Nxc8 Qxc8 21.Bd3 Be8 22.Be4 a5 23.a4
White gains nothing from pushing the f pawn: 23.f5 Bc6 24.Bg5 Qd7 25.Qe2 Bxe4 26.Bxe7 Qxe7 27.Qxe4 Rae8 28.Rae1 Qg5. Granted, the position is still equal here.
23...Ra7 24.Qe1 b6 25.Qg3 Bg6 26.Bg2 Be8 27.Be4 Bg6 28.Bg2
White begins to repeat moves, but Baadur sees it as a signal to start the fight. About time!
28...Bc2 29.Rae1 Qe8
The queen obviously heads for g6.
The computer believes that White can just assume a wait-and-see approach: 30.Rf2 Qg6 31.Bc1 d3 32.Rd2 (32.f5 exf5 33.Ref1) 32...h5 33.Be4 with equality, but this breakthrough is in the spirit of White's game and is objectively good, with Black facing more risks.
Now Black has to consider the possible intrusion of the white queen on d6 on every move.
If 31...Qg6, the possible response is 32.Rc1 d3 33.Qd6, as well as the simple 32.h3.
Nothing can be gained from 32.Qd6 Qd8.
White can have some fun by making the planned move immediately: 32.Rf2!? fxg4 33.Rf7! At the end of the day, the position is still equal: 33...Bg6 34.Rxf8+ Qxf8 35.Qd6 Qc8 36.Qxb6 Ra6 37.Qb5 Rxe6 38.Rxe6 Qxe6 39.Qxc5 Qe2 40.Qxd4 Qd1+ 41.Kf2 Qxb3ч 42.Qd8+ Kf7 43.Bd5+ Nxd5 44.Qxd5+ Kf8 45.Qd6+. But Vachier-Lagrave makes a natural human move and protects the pawn.
32...Kh7 33.Rf2 fxg4 34.Rxf8 Qxf8 35.hxg4 Qd8
The balance is now in White's favor. Black should have considered something like 35...Qe8 36.Rc1 Qg6 37.Qb8 Be4 38.Bxe4 Qxe4 39.Qxa7 Qxg4+ with a perpetual check. After the queens are traded off, the pawn becomes even more dangerous.
36.Rf1 Bg6 37.Qe5 Qc7 38.Bf4 Qxe5 39.Bxe5 Ng8 40.Bd6?
It seems that White has missed a good chance of winning. It's not clear how to defend after 40.Rf8! Nf6 (or 40...Re7 41.Bd5 Nf6 42.Bd6 Ra7 43.Rd8) 41.Rd8! Re7 42.Bxf6 gxf6 43.Rd7 Rg7 44.Bc6, and White wins a bishop for his passed pawn. In most lines, Black's passer is successfully stopped by the white king.
40...Nf6 41.Bc6 Nxg4 42.e7
Once again, White abstains from placing his rook on the eighth rank, and once again, it's a mistake. After 42.Rf8 d3 43.Bf3 Nf6 44.Kf2 h5, Black's extra pawns are only enough to make up for his own passive pieces, for example: 45.Bh1 h4 46.Ke3 Ne8 47.Be5 Re7 48.Bd5 Nc7 49.Bxc7 Rxc7 50.Rf4 Kh6 51.Rxh4+ Kg5 52.Rh8 Bf5 53.Rb8 g6 54.Rxb6 Rh7 with equality. In the game, however, Black manages to gain an edge: having extra material is no laughing matter.
42...Nf6 43.Rf4 d3 44.Kf2 Ne8 45.Bxe8 Bxe8 46.Rf8 d2 47.Ke2 Rd7 48.Rxe8 d1Q+ 49.Kxd1 Rxd6+ 50.Ke2 Re6+ 51.Kf3 g5 52.Rb8 Rxe7 53.Rxb6 h5 54.Rc6 Rf7+ 55.Kg2
The connected pawns should bring Black a win, but he has to make his king active first. This goal should have been attained with 55...Rf5! 56.Kh3 (56.Rc7+ Kg6 57.Rc6+ Kf7 58.Rc7+ Ke6 59.Rc6+ Ke5 60.Rxc5+ Kf4) 56...Rf3+ 57.Kg2 Rxb3 58.Rxc5 Kg6 59.Rxa5 Rc3 with a win.
55...h4? 56.Rxc5 Kg6 57.Rxa5 g4 58.Ra8 Rf3 59.Rg8+ Kf5 60.Rf8+ Ke4
Black activated his king, but...
61.Re8+ Kf4 62.Rf8+ Ke3 63.Re8+ Kd2 64.Rg8 h3+ 65.Kh2 Rf2+ 66.Kh1 Rf1+ 67.Kh2 Rf2+ 68.Kh1 Draw.
Dominguez – Jakovenko
1. d4 Nf6 2.c4 e6 3.Nf3 d5 4.Nc3 c6 5.g3 dxc4 6.Bg2 Nbd7 7.a4 Bb4 8.0–0 0–0 9.Qc2 Qe7 10.a5 e5 11.Rd1 a6 12.d5 Nc5 13.Be3
In a recent game Giri – Ivanchuk, Wijk aan Zee 2015, the opponents played 13.Bg5 Bxc3 14.bxc3 Bg4 15.h3 Bxf3 16.Bxf3 e4 with unclear play.
13...Nb3 14.Ra4 c5
The line 14...e4 15.Ng5 Bxc3 16.d6 Qe5 17.Qxc3 Qxc3 18.bxc3 Nd5 needs checking; now White has to choose between having a pretty draw 19.Bxe4 Nxc3 20.Rxc4 Nxd1 21.Bxh7+ Kh8 22.Bc2! (despite having a rook up, Black has to content himself with a draw: 22...Nxe3 23.Rh4+) and continuing the fight with unclear consequences: 19.Nxe4. The move in the game seems to give White an advantage.
15.d6 Qe8 16.Rxb4 cxb4 17.Nd5 Nxd5 18.Rxd5 Nd4
One important option is 18...e4 19.Ne5 Be6 20.Bxe4! with a major edge for White, since 20...Bxd5 21.Bxh7+ Kh8 22.Qf5 loses on the spot.
The knight on b3 becomes Black's headache in the line 18...h6 19.Nxe5 Be6 20.d7 Qe7 21.Nxc4 Bxd5 22.Bxd5 Qxd7 23.Nb6 Na1 24.Qb1, but not for a long time: until its elimination on the next move.
This capture proves to be inaccurate and allows Black to equalize the game. After the correct 19.Nxd4 exd4 20.Rxd4 Be6 21.Bxb7 White has an advantage because 21...Ra7? is bad due to 22.Rh4.
19...exd4 20.Rxd4 b3 21.Qxc4 Be6 22.Qd3 Qc6 23.Qd2 Rac8
Interestingly, 23...Qc2 would have been pretty good now, while the queen's intrusion along the c file on the next move will be a mistake.
This resulted in an ultimate win for White, but an objectively better option would have been 24.d7 Rcd8 25.Ng5 Qc2 26.Nxe6 Qxd2 27.Rxd2 fxe6 28.Bxb7 Rf5 29.Bxa6 Rxa5 30.Bc4 with a probable draw.
Black underestimates White's 26th move and chooses a plan that results in a disaster. At the same time, after the correct 24...Qb5 25.f4 Rfd8, preparing Rc4, Black would have had a considerable advantage.
It was not too late to change plans: after 25...Qxd2 26.Rxd2 Rc5 27.Bxb7 Rxa5 28.e4 g6 the plan Ra1 with the following a6–a5–a4–a3 ensured excellent prospects for Black.
26.Kf2! Rc2 27.Qb4 Re8
27...Bg4 28.Bf3! doesn't help either.
The only move leading to a win, but it suffices.
No better is 28...Bd7 29.Qxb7 Rxe2+ 30.Kxe2 Qxb2+ 31.Rd2 Qc3 32.Qxd7 Rxe4+ 33.Kf1, and White wins.
29.Nxc2 bxc2 30.Qa4 Kf8 31.d7 Bxd7 32.Rxd7 c1Q 33.Qa3+ Kg8 34.Bxb1 g6 35.Qb3 Qc5+ 36.e3 Black resigns.
Tomashevsky – Grischuk
1. d4 Nf6 2.c4 g6 3.Nc3 Bg7 4.e4 d6 5.Nf3 0–0 6.h3 e5 7.d5 Nh5 8.g3 Qe7
An opening subtlety. Black probably hopes to exploit the opposition of his queen and the White king in certain lines. The drawback of the move is that e7 is usually a bad square for the queen in the King's Indian. Given Black's next move, it seems appropriate to show what happens after the immediate 8...Na6: 9.Be3 f5 10.exf5 gxf5 11.Ng5 (11.Nh4 f4) 11...Qe8 12.Be2 Nf6 13.Qd2 Qg6 14.0–0–0 Bd7 15.h4, Karyakin – Nakamura, Tashkent 2014, with unclear play.
Black's idea will probably work after 9.Be3 f5 10.exf5 gxf5 11.Ng5!?, but now after 11...Nxg3 12.fxg3 f4 13.Qh5 h6 14.Nge4 fxe3 15.Qe2 White is still better. 11...Nf6 is a preferable option, where in some lines Black is ready to capture the pawn on e6. I can't boast that I have sorted all this out. White chooses a simple solution: preventing f7-f5 and continuing development. Black has to lose tempos and, in some lines, suffers from the awkward position of his queen.
9...Na6 10.Be3 Nc5 11.h4 a5 12.Be2 Nf6 13.Qc2 c6 14.g4 Na6
Now a5-a4 and Qa5 are impossible, and Black brings the knight back because after the immediate 14...cxd5 there is the unpleasant 15.Bxc5 (15.cxd5 Bd7 16.g5 Ne8), and now both after 15...d4 16.g5 dxc3 17.gxf6 Qxf6 18.Be3 cxb2 19.Rb1 Qxh4 20.Rxb2І, and after 15...dxc5 16.g5 d4 17.gxf6 Qxf6 18.Nd5 Qxh4 19.0–0–0 (there is no 19...Qxf2?? 20.Nf3) White's piece is stronger than the three pawns.
15...Nh5?! 16.Bxh5 gxh5 17.Qe2 is bad: Black has no compensation.
A different placement of the knights is no better: 15...Nd7 16.h5 Ndc5 17.Qd2 cxd5 18.Nxd5 Qd8 19.f3 Nb4 20.Rd1 Nc6 21.Nf6+ Bxf6 22.gxf6 Qxf6 23.Qxd6.
16.h5 Nb4 17.Qd2 cxd5 18.cxd5 Bd7 19.0–0–0 f5 20.Kb1 b5 21.hxg6 hxg6
22.a3, with the aim of capturing the pawn after the knight's withdrawal, seems tempting, but the knight is not going anywhere: 22...Qd8 (both 22...f4 23.Bb6 Ra6 24.Bxa5 Rxa5 25.Na2! and 22...Na6 23.Nxb5 are bad) 23.f3 (23.Nf3) 23...Nc7. The threats along the h file probably give White an edge after a move like 24.Qe1, but there is no certainty, while chances for an attack are already created.
Another idea is also interesting: 22.Nf3!? f4 (the attempt to make the position more active 22...fxe4 23.Nh4 Qf7 24.a3 Nd3 25.Nxe4 Nf4 26.Bxf4 Qxf4 27.Qd3 Bf5 28.f3 b4 29.Ng2 Bxe4 30.fxe4 Qxg5 31.Rdg1 ultimately loses) 23.Nh4! fxe3 24.Qxe3 Rf4 (24...Qf7 25.Nxg6 Qxg6 26.Bh5 Nxd5 (26...Qh7 27.Bxe8 Qxh1 28.Rxh1 Rfxe8 29.Qb6, and, if what the computer says is true, the white queen is better at taking the pawns then the black rooks, and the two minor pieces manage to work together) 27.Nd5 Qe6 28.Bg6!) 25.Nxg6 Qxg5 26.Rdg1 Qf6 27.Rh8+ Bxh8 28.Nxf4+ Bg7 29.Nh5 Qh4 30.Qf3. These are promising lines, even though they are not completely clear. This is why White makes a logical consolidating move without giving Black any extra chances to complicate the game.
22. f3 Nc7 23.Rc1 Nba6 24.Bd3 b4 25.Ne2 Nb5 26.Rc6! Rfb8 27.Rhc1 f4 28.Bf2
Before this, White acted very logically, even though a superficial analysis can hardly show whether his actions were perfect. Black's prospects look quite gloomy and, given that starting from this moment his position only got from bad to worse, it seems a good idea to check whether he should have grabbed the rook out of sheer desperation: 28...Bxc6 29.dxc6 Nac7 30.Ng4 Ra6 31.Bc4+ (Black plans Qe8, for example: 31.Ka1 Qe8 32.Bc4+ Kf8 33.Nxf4 Rxc6 or 31.Bh4 Qe8) 31...Kf8 32.Nxf4! Such a strike seems to be decisive, but it's not that clear: 32...exf4 33.Qxf4+ Ke8 34.Nf6+ Kd8 35.Nd7 Rba8 36.Bb6 Rxb6 37.Nxb6 Rb8 38.Nd7 Rc8 39.Nf6 Nd4 40.Rd1 Ncb5 – miraculously, Black manages to survive. I am not sure that White doesn't have any decisive attack, but Black has nothing to lose anyway.
29.Ng4 Qxg5 30.Rh1 Qe7 31.Bxb5 Rxb5 32.Nxf4! Bxg4
Or, as I will point out just in case, 32...exf4 33.Bd4 Bg7 34.Nh6+ Kf8 35.Qxf4+ Ke8 36.Bxg7 Qxg7 37.Nf7 Bxc6 38.Rh8+ Qxh8 39.Nxh8 with a victory.
33.Nxg6 Qf6 34.Nxf8 Rxf8 35.Rg1
Some Bulgarian cheater could go for 35.Rxd6! Qxd6 36.Qg5+ Kf7 37.Rh7+ Ke8 38.Qxg4.
35...Qxf3 36.Qg5+ Kf7 37.Rxg4 Ke8
It's clear that the check doesn't help: 37...Qd3+ 38.Rc2.
It's difficult to avoid seeing nightmares in this position given that the move 39.Qxf7+?! was made.
Many of us have made moves like that seeing that the flag is about to fall. None could explain later why they made them. Evidently, 39.Qxd6, и 39.Rc8+ Kd7 40.Rxf8 Qxf8 41.Qh7+, etc. is not bad either.
Nevertheless, it was this move that was the real blunder. If it had been move 41, the game would have lasted just another two moves: 40.Rg8+ Kd7 41.Rxa6.
Some more drama is added: Black misses a certain draw 40...Rf1+ 41.Rc1 (41.Kc2 b3+!) 41...Rxc1+ 42.Kxc1 b3 43.a3 Nc5 and has a losing position once again.
41.Rxd6 Rf3 is not so clear.
41...Kd7 42.Ra8 Rb7 43.Bxc5
43.Bg5 Rf1+ 44.Kc2 Rf2+ 45.Kd1 Ne6 46.dxe6+ Kxc6 47.e7 Rxe7 48.Bxe7 Rxb2 49.Rxa5 b3 50.a3 would have been sufficient for a win: Black can't trade off all the pawns. But the move in the game looks, and probably is, simpler.
I suspect that White made a mistake and took the wrong pawn. He had a sure win after 44.Rxa5 Rf1+ 45.Kc2 c4 46.Rh6 (или 46.Rxc4 b3+ 47.axb3 Rf2+ 48.Kb1 Rxb3 49.Rc2) 46...Rf2+ 47.Kc1 Rf1+ 48.Kd2 Rc7 49.Ke2 (49.Rh7+ Kc8 50.Rxc7+ Kxc7 51.Rc5+ Kb6 52.Rxc4 Rf2+ 53.Kc1 Kb5) 49...Rf4 50.Ke3 c3 51.Rh7+ Kc8 52.Rxc7+ Kxc7 53.bxc3 bxc3 54.Rc5+ Kd7 55.Rxc3.
The cleanest way to win was 44.b3! Rf1+ 45.Kc2 Rf2+ 46.Kd3 thanks to the idea 46...Rxa2? 47.Rh8.
White probably missed 45.Kc2 b3+.
45...Rxc1+ 46.Kxc1 b3! 47.a3 Rc7+ 48.Kd2 Rc2+ 49.Kd3 Rxb2 50.Rxa5 Rh2 51.Rb5 Rh3+ 52.Kd2 Kd6 53.a4 Rh2+ 54.Kc3 Rh3+ 55.Kd2 Rh2+ 56.Kc3 Rh3+ 57.Kc4 Re3 58.Rb6+ Kc7 59.Rxb3 Rxe4+ 60.Kc5 Rxa4 61.d6+ Kd7 62.Rb7+ Ke6 63.Re7+ Kf5 64.d7 Rd4 65.Kc6 e4 66.Re8 Kf4 67.d8QRxd8 68.Rxd8 e3 69.Kd5 e2 70.Re8 Kf3 71.Kd4 Kf2 72.Kd3 e1Q 73.Rxe1 Kxe1 74.Kd4 Draw
Giri – Karjakin
1.Nf3 Nf6 2.g3 c5 3.Bg2 Nc6 4.0–0 e5 5.e4 Be7 6.Nc3 d6 7.d3 0–0 8.h3 Rb8 9.Nh2 Bd7 10.Ng4 Qc8 11.Nxf6+ Bxf6 12.Kh2 Be6 13.f4 exf4 14.gxf4
This is clearly the only way and that's how Black planned to defend from the push f2-f4, but still... it's always a pleasure to see a maneuver from the old chess legacy being executed.
15.bxc3 f5 16.Be3?!
I think that White's problems are largely connected with this move and the bishop's subsequent fate. Unlike in a classical King's Indian game, where Black captures with the "King's Indian bishop" on c3, White still has hopes to activate the bishop along the long diagonal, and that's how he should have blackmailed Black: 16.Rg1.
Further on, Black acted in a strong and logical manner and nothing peculiar happened... until the last move.
16...Qc7 17.Qd2 fxe4 18.dxe4 Na5 19.Bf2 Nc4 20.Qc1 Bd7 21.f5 Bc6 22.Bh4 Rbe8 23.Qf4 Qa5 24.f6 Rf7 25.Qg3 Nd2 26.Rf4 Nxe4 27.Bxe4 Rxe4
27...Bxe4 seems to be better.
28.Rxe4 Bxe4 29.Re1 Bc6 30.Qxd6
Now after the capture 30...Qxa2! White has serious problems: 31.Re2 (after 31.Qd8+ Rf8 32.Qd2 Qd5 33.Qxd5+ Bxd5 34.Re5 Bc6 White still has to struggle for a draw) 31...Qc4, and now the following amusing line is possible: 32.Re7 Rxe7 33.Qxe7 Qf4+ 34.Kg1 Qc1+ 35.Kf2 Qxc2+ 36.Kg3 Qxc3+ 37.Kf2 Qd2+ 38.Kg3 Qg2+ 39.Kf4 Qf3+ 40.Ke5 Qe4+ 41.Kd6 Qxe7+ 42.Kxe7 gxf6 43.Bxf6; it's highly likely that the three extra pawns won't be sufficient for a win.
But when our king is threatened, decisions are often made that we later regret.
31...Qc7 31.Qxc7 Draw.