26 May 2015

When You Fail to Score You Get Scored On

Round Three analysis of the FIDE Grand Prix in Khanty-Mansiysk by GM Sergei Shipov.

The Americans demonstrated a brilliant performance in the first two periods against the Russian team in the Semifinal of the World Hockey Championship in Prague. They had a lot of opportunities but failed to score. And in the third period hockey gods punished them for it!

Just shortly before that their fellow chess Olympians punished Tomashevsky, who, if you still remember it, had failed to make use of a whole bunch of scoring chances against Grischuk in the second round of this Grand Prix.
And god's punishment came down in the person of the rating favorite of the tournament.

I think that the fatigue has taken its toll on the grandmaster from Saratov. He will paint complex strategic pictures in each and every game of his. Extending oneself like this into the game on a daily basis must be so incredibly exhausting...

Caruana – Tomashevsky 

1.d4 d5 2.c4 c6 3.Nf3 Nf6 4.Nc3 a6 5.a4 e6 6.Bf4!? a5 7.e3 Be7 


This is a brave novelty and quite a logical one. The weakness of the g7-pawn and the overall passivity of Black’s setup provide White with fertile ground for drastic actions on the kingside.

Relatively quiet continuations 8.Bd3, 8.Be2 and 8.h3 have been tested in this position before. 


The idea of launching a reciprocal attack with 8...h6!? 9.Rg1 g5 10.Bg3 Nbd7 etc. is not without interest. 

9.g5 Nd7 

Not good is 9...Nh5 10.Be5 Bxg5? (or 10...f6 11.gxf6 gxf6 12.Nh4 Ng7 13.Bf4±; 10...0–0 11.Nd2 g6 12.Be2 Ng7 13.h4 f6 14.gxf6 Bxf6 15.Nf3 b6 16.Qd2+/-) 11.Nxg5 Qxg5 12.h4 Qh6 13.Be2 Nf6 14.Bf4 Qg6 15.h5 Qf5 16.Bd3 Qg4 17.f3! Qg2 18.Rh2 Qg1+ 19.Bf1, and the threat of Nc3-e2 cannot be fended off. The wild queen will be captured inevitably. 

10.h4 Nb4 11.Be2 b6 

An attempt to offer a pawn as a sacrifice to create counterplay via 11...0–0 12.h5 e5!? is going to backfire. After 13.dxe5 dxc4 14.Bxc4 Nc5 15.Qxd8 Rxd8 16.Ke2 Bf5 17.g6! hxg6 18.Rag1 it is Black who is at the receiving end of an attack. 

12.h5 Bb7 

12...Ba6!? offered chances for equality as well, although in order to achieve this purpose a great deal of precision would be required: 13.cxd5 (or 13.b3 Rc8 followed by с6-с5) 13...Bxe2 14.Qxe2 exd5 15.e4 0–0 16.0–0–0 Qc8! (16...Re8 17.g6!) 17.exd5 (17.Kb1 dxe4! 18.Qxe4 Re8) 17...Re8 18.g6 Bf6 19.Qd2 cxd5 20.Kb1 Nf8 21.gxf7+ Kxf7 with equality. 


Neither during online broadcasting nor in my home analysis was I able to comprehend the profound idea behind the computer’s recommendation 13.Kf1!? So be it! 


On the other hand, after 13...cxd5 14.Nb5 Rc8 the idea behind the move 15.Kf1! becomes quite clear as we are stepping out of the knight’s fork while avoiding the unnecessary trade of material – 15...0–0 16.Nd6 Bxd6 17.Bxd6 Re8 18.Rg1 Ba6 19.Bb5 Bxb5+ 20.axb5 Rc4 21.Rg3 Nb8 22.b3 Rc3 23.Bxb4 axb4 24.Ne5 with an edge for White. 




This is not the best continuation, to my mind. The regrouping of pieces with 14 ... Nf8! would have been a lot more reliable, for example, 15.0-0-0 Ne6 16.g6 fxg6 17.hxg6 Nxf4 18.exf4 h6 19.Ne5 0-0 20.Nf7 (20.Bg4 Bc8) 20 ... Rxf7 21.gxf7 + Kxf7. This position was the object of mutual discussion with Sergei Rublevsky in our online broadcast. I keep standing pat on my point of view that Black’s position is OK.


15.Bd3! 0-0 16.0-0-0 looked rather unpleasant for Black as White is ready to exert additional pressure on the f5-pawn by means of Nf3-h4. No optimistic scenarios have been found: 16 ... Nxd3 + (16 ... b5 17.Bb1 bxa4 18.Nh4 Bxg5 19.Bxg5 Qxg5 20.Rdg1 with the attack; 16 ... Qe8 17.Nh4 g6 18.hxg6 Nxd3 + 19. Qxd3 hxg6 20.Nxg6 Qxg6 21.Rh6 Qg7 22.g6 and White has the advantage) 17.Qxd3 Ba6 18.Qc2 b5 19.g6 h6 20.axb5 Bxb5 21.Nxb5 cxb5 22.Kb1 with better chances for White

15...Bxf6 16.h6 g6 17.e4 dxe4? 

This is a serious positional concession. There is no justifying the idea of surrendering the center to the opponent. Stronger is 17...0–0!, and, for example, after 18.e5 Be7 19.0–0–0 c5! and Black has counterplay. 

18.Nxe4 0–0 

From now on the only item on the agenda is about White’s choosing from multiple attacking possibilities in terms of their relative success rate. 


More precise would be 19.Bc4+! Nd5 (19...Kh8 20.0–0!?) 20.0–0–0 Kh8 (20...b5 21.Ba2 Be7 22.Ne5 with advantage for White) 21.Nd6 Ba6 22.Bxd5 cxd5 23.Kb1 Be7 24.Ng5 with an edge for White. 


19...b5! 20.Rhe1 Be7 21.Bg3 (21.Neg5 Nb6 22.Ne6 Qd5) 21...Nb6 22.Ne5 Bc8! was an interesting option here followed by the transfer of this bishop on f5. However, such tightrope-walking is almost impossible to be performed during the game. 

20.Bg3 Ba6 

At this moment the computer comes up with the trick of 20...b5 21.Nfg5 N7b6 with the idea of 22.Ne6?! Qe7 23.Nxf8? Qxe4, promising comfortable position for Black. 

However, stronger is 22.Nxf6+ Qxf6 23.Be5 Qe7 24.Rhg1! and White’s threats are very dangerous. 

21.Bxa6 Rxa6 22.Rhe1 

More precise would have probably been to play 22.Qc2! Ra8 (22...b5 23.Ne5) 23.Kb1!, when the weakness of the с6-pawn prevents Black from successfully organizing his defensive activities: 23...Rc8 24.Nd6 Rc7 25.Rhe1, etc. 

22...Ra8 23.Kb1 Be7 24.Qd3 Rf5 (24...Rc8!?) 25.Ne5 Qc8 

More stubborn is 25...Nxe5 26.dxe5 Qd7. 


This move is aimed at removing the last obstacle to White’s attacks in the center.


Or 26...Nxe5 27.Nxd5! cxd5 28.dxe5 with an edge for White. 

27.Qxc3 Nxe5 28.Qb3+ 

28.dxe5 Qe6 29.Rd6! Bxd6 30.exd6 Qf6 31.Qxc6 Rf8 32.Re7+- was no less spectacular a win. 


28...Nf7 29.Rxe7 Qf8 30.Rde1 Qxh6 31.Qxb6+/- was no piece of cake either. 

29.dxe5 Qf5+ 30.Ka2 Bb4 

Some practical survival chances in the position still remained after 30...Qg4! 31.e6 Rf3, although it must be said that they were rather slim ones: 32.Qc2 Qb4 33.Re4 Qc5 34.Rc4 Qf5 35.Qxf5 Rxf5 36.Rxc6+-. 

31.e6! Re7 32.Bh4 Ree8 (32...g5 33.Bxg5 Qxg5 34.Rg1+-) 33.e7+ Qf7 



This is a mighty blow! 

34...b5 35.Rd8 bxa4 36.Qe3 

Here we analyzed also 36.Qc4 followed by 36...Bxe7, but it also loses to: 37.Rd7! Rad8 38.Rxg6+! hxg6 39.h7+ Kg7 40.h8Q+ Rxh8 41.Rxe7+-. 

36...Bxe7 37.Rxa8 Rxa8 38.Bxe7 Re8 39.Ka1 a3 40.bxa3 Qf5 41.Qc3 Black resigs.
This was a natural result of the game that was won by a player who was being stronger and less exhausted. Here and now.

Yet another Russians player, who shone at the start, had his game going poorly for him. Therefore, his fellow compatriot triumphed in the game.
Someone person's loss is another person's gain. Someone comes, and someone goes...

Jakovenko – Svidler 

Being somewhat disappointed by the unhappy outcome of the middle-game, Dmitry failed to put up an adequate resistance in the ending that featured real drawing chances. 


This is a rather passive play. 

Indeed, White would be OK in the pawn endgame. However, the rook should be handled in a more energetic manner: 43.Rd4!, for example, 43...b5 44.Re4+ Kd6 (44...Re5 45.Rxe5+ Kxe5 46.h4=) 45.h4 Rd5 46.Rg4 g5 47.hxg5 (47.h5 Rd3+ 48.Ke2 Ke5 49.Rd4 Rxd4 50.cxd4+ Ke6 51.g4 a5–+) 47...fxg5 48.Re4 with a defendable position. 

43...b5 44.Rd2 Rd5 45.Re2? 

This is already a blunder. Rook activation was already a must. Here is an approximate line: 45.Rd4! f5 46.Rh4 Rd3+ 47.Ke2 Ke5 48.Rh8 Kf4 49.h4 Rd6 (or 49...Kg4 50.h5 Kg3 51.Rg8) 50.Rh7 g6 51.Rh6 Kg4 52.h5 Kg5 53.Rxg6+ Rxg6 54.hxg6 Kxg6, when for yet another time the arising pawn ending does not favor Black decisively: 55.Ke3 Kg5 56.Kf3 c5 57.g3 a5 58.Kf2 Kg4 59.Kg2 b4 60.a4=. 

45...Kf5! 46.Rf2+ 

No better is 46.Kf3 Rd3+ 47.Kf2 Kf4! 

46...Ke5 47.Kf3 Rd3+ 48.Kg4 

48.Ke2 Ke4–+. 

Black has managed to squeeze White out of the center in an amazingly prompt an easy fashion. Now his penetration into the queenside has been facilitated. 


A lot simpler is 48...Ke4! 49.Kh5 Ke3 50.Rf3+ Kd2 etc. 


49.Kh5 is most precisely answered with 49...Rg3 50.Kh4 Rg5 51.g4 Ke4 52.Rd2 Rd5 53.Re2+ Kd3 54.Rh2 Ke3 (54...a5) 55.Kg3 Rd2–+. 

49...Ke4 50.Rf2 



This uncomplicated breakthrough finishes the game. 


51.Kh5 Rg3 52.Kh4 f4–+ is simply terrible for White. 

51...Re2 52.g3 Rxb2 53.Rg5 Kd3 54.Rxg7 Kxc3 55.Rc7 a5 56.Rxc6 b4 57.axb4 axb4 

There is a great difference between White’s and Black’s passed pawns. White resigns. 

So far we have failed to see the energetic and motivated type of play from one of the tournament favorites. 

Grischuk – Vachier-Lagrave 

1.e4 c5 2.Nf3 d6 3.d4 cxd4 4.Nxd4 Nf6 5.f3 e5 6.Nb3 d5 7.Bg5 d4 8.c3 Nc6 

This variation is evaluated as not an easy one for Black. This is yet another game that has failed to change this settled evaluation. 


It is also not clear how Black is supposed to equalize after 9. Bb5! However, the answer to this question is still hidden behind a veil of obscurity... 

9...exd4 10.Bb5 Bb4+ 11.Bd2 Bd6 12.Na3 

12.Bxc6+ bxc6 13.Ba5! with the intention of capturing the d4-pawn with the queen looked quite smart. 

12...0–0 13.Nc4 Be6 

In this position Black could have already tried to claim more with 13...Bc7! 

14.Nxd6 Qxd6 15.0–0 a6 16.Bxc6 bxc6 17.Rc1.
At first I believed that White has the upper hand after 17.Qc1 Bxb3 18.axb3 c5 19.Qc4 Nd7 20.b4, but, upon paying additional attention to the 20...Rfc8 continuation, I failed to come up with any constructive ideas for White on how to improve his position further. 


Black has demonstrated his intention to carry out the с6-с5 advance. However, White’s combination deprived Black of this plan ... 

18.Nc5 Nxc5 19.Bb4 

...only to find out that the resulting opposite-colored bishop middle-game offered him no winning chances. 

19...Rfd8 20.Bxc5 Qb8 21.b3 a5 22.Rf2 Qb5 23.Rd2 a4! 24.Bxd4 axb3 25.axb3 Bxb3 26.Qe1 f6 27.Qe3 Ra2 28.Rxa2 Bxa2 29.Qa3 Bf7 30.h3 h6 31.Qc3 Qa4 32.Bf2 Rd1+ 33.Rxd1 Qxd1+ 34.Kh2 Qd6+ 35.Bg3 Qe6 36.Qd4 Kh7 37.Bd6 Qd7 

38.f4 f5 39.exf5 Qxf5 40.Be5 Bd5 41.Qa7 Qf7 42.Qxf7 Draw. 

Another compatriot player of ours, Sergei Karjakin, will not have his game going smoothly for him. Every encounter sees him struggling in the time trouble, where the quality of his performance comes down. But so far fortune has favored him. 

Karjakin – Dominguez 

At this moment Leinier had 20 minutes on his clock, whereas Sergei had only 5! In addition to that, Black’s position is better. 

Firstly, Black had an option of playing up a pawn via 30...Bxf3 31.Bxf3 g5! 32.hxg6 (32.Bg3 Bg7) 32...Nxg6 33.Bg3 Ndxe5 34.Be4 (34.Bg2 Bg7) 34...Qb4 – with winning chances. 

Secondly, the retreat of the bishop was connected with the subsequent threat of the queen penetration into White’s camp and the activation of his knights. 
Not so convincing in this respect would be to go for 30...Bh7 in view of 31.g5 Nf5 32.gxh6 gxh6 33.Qd2. 

On the other hand, a simple 30...Bc6 would be stronger, for example, 31.Bg3 (31.Ng3 g5!; 31.g5 Bxf3!) 31...Qb1 32.Kh2 Nb6 33.Ne3, and now 33...Ba8!? with the idea Ne7-c6-d4 looked quite interesting. 

By the way, it gives you the idea of an immediate retreat of the bishop to a8... 

In general, Black could and should have played for a win since he was not without reasonable chances of success. 

In the game, however, Dominguez retreated his bishop to с6 and offered a draw. Karjakin, of course, had no objection to this. 

Why did it happen the way it happened? To my mind, the good tournament standing of the Cuban grandmaster is to be blamed. Had he had fewer points he would have striven very much to increase their quantity.  But the way it is... 

And, ultimately, let me present you the best game of the third round for dessert. 

Jobava – Gelfand 

1.d4 Nf6 2.c4 g6 3.Nc3 d5 4.Nf3 Bg7 5.Qa4+ Bd7 

Here everyone without exception would retreat his queen to b3; the rebellious Georgian undergrounder, however, did find his own way. 


This is the latest trend in the theory of this variation. The queen guards the с5-square while exerting pressure on the e7-pawn. On the other hand, this is a true pawn sacrifice. Boris chose the most principled continuation without wasting time for extensive contemplations. 

6...dxc4! 7.e4 b5! 8.e5 Ng4 

At first the computer evaluates the retreat of the knight 8...Nh5!? to be a decent continuation. The point is that White has no time to trap this knight: 9.h3?! a5! 10.Ne2 Na6! 11.g4 (11.Bd2 Bc6 12.g4 Nb4 13.Bxb4 axb4 14.Qe3 Bh6! does not offer White better prospects either) 11...Nb4 12.Qc3 Nd3+, and White is in a very bad shape. 

However, the position of the knight on the rim of the board is going to tell in the following continuation:  9.Be2! a5 10.Nxb5! Bxb5 11.g4 with a small edge for White. 


Jobava made this move rather fast, demonstrating that he was still within his home preparation.

White will not profit from 9.Nd5 in view of 9...Nc6 (9...Bc6?! 10.Nf4!) 10.h3 (10.Qc5 e6) 10...e6! (rather than 10...Nh6? 11.Qc5! e6? 12.Bxh6 Bxh6 13.Nf6+, and White is winning) 11.Bg5 f6 with unclear complications. 


The tempting 9...a5 would be refuted by trading of material via 10.Nxb5! Nxf2 11.Nxc7+ Qxc7 12.Kxf2 – when the consolidation of the d4-square is going to be followed by exerting further pressure on the с4-pawn. 


In one of his future games in this line Baadur is very likely to choose 10.g4!? a5 11.Ne2 (11.Qc5? Na6 12.Qd5 Nb4! is, of course, not good) 11...Nc6 12.Bd2! (12.Bg2?! Nb4) 12...0–0 13.Bg2 f6 14.Qc5 Nb4 15.Bxb4 axb4 16.0–0 with a rather unclear and complicated game. 

So, we strongly advise the proponents of the Grunfeld defense to start preparing without further delay! 


The inherent drawback of White’s queen position on a3 starts to tell. Black pawns are advancing with tempos. 10...cxb3? would fail to 11.Bxb5+/-. 


11.Bf4 b4 12.Qc1 suggested itself, but the subsequent trade of material favored Black: 12...cxb3! 13.axb3 (bad is 13.Bxh6? Bxh6 14.Qxh6 b2–+, or 13.Ne4 Bc6 14.Bd3 Nf5 15.Qc5 Qd5!) 13...Nf5 14.g4 (14.Na4 0–0 15.Qd2 Bc6!) 14...Bc6 15.d5 Bxd5 16.Nxd5 Qxd5 17.gxf5 Qxf3 18.Rg1 Qe4+ 19.Qe3 Qxf5 20.Rc1 0–0 21.h4 Qd7 22.h5 a4-/+. 

11...b4 12.Ne4 

It took Gelfand long time to decide on his next move and it looks to me that I managed to guess at the reason behind it. 

It makes a lot of sense to start undermining White’s center immediately: 12...Bc6!?, for example, 13.Qc2 Nf5 –and now, if compared to what happened in the game, White is devoid of the consolidation maneuver  Qc2-d3. One of the possible continuations was 14.Bxc4 (14.Be3!? c3 15.Bc4) 14...Nxd4 15.Qd3 Nxf3+ 16.Qxf3 0–0 17.Bb2 Nd7 18.0–0–0 Qc8 19.e6 Bxb2+ 20.Kxb2 Nb6 21.exf7+ Kg7, and Black’s chances should be preferred. 

12...c3 13.Qc2 

The protected passed pawn on с3 is no warranty of a comfortable life for Black as he may fail to last as long as to see the endgame. 


In the post-game analysis both players agreed that this was a hasty decision. However, the analysis proves that the issue is not that simple. After 13...Bc6 14.Bc4 Black is up against a choice. In the case of 14...Nf5 15.Qd3 it is not so easy to come up with a good continuation for Black: 15...Nd7? (15...Bd5 16.g4!; 15...e6 16.Bg5) 16.Bxf7+! Kxf7 17.Qc4+. On the other hand in case of 14...Nd7 15.d5 Bb7 16.e6 Nb6 17.Bxh6 Bxh6 18.Nfg5 Bxg5 19.Nxg5 0–0 20.exf7+ Rxf7 21.Nxf7 Kxf7 22.0–0–0 Black has a definite amount of compensation for the exchanged sacrifice, but no more than that. 

14.Bc4 Bc6 15.h4! 

This is the point. White’s intended assault along the h-file is going to be extremely dangerous for the black king. White would gain nothing from 15.Bxh6 Bxh6 16.h4 Kg7 17.h5 Bd5, to be followed with Nb8-c6 and Black is better. 

15...Nf5 16.Qd3 

I can recommend meeting 16.Be3 with 16...h5! as the direct assault by White is going to fail in this case:  17.e6 f6 18.Ng3, and now not 18...Nxg3?! in view of 19.Qxg6! Be4 20.Qxg3 Kh8 21.Qf4 Bg6 22.d5 Qd6 23.Nd4 with advantage for White, but 18...Nxe3! 19.fxe3 f5 with better perspectives for Black. 


In the case of 16...h5 17.e6 17...f6 was no longer good as it failed to 18.d5+/-. So, Black is forced to play 17...fxe6!, although such weakening of one’s position is very hard to agree to. However, my analysis proves advantage for Black after 18.Bxe6+ Kh8 19.d5?! (19.Be3 Bd5 20.Nfg5 Nc6 21.Ng3 Bxe6! 22.Nxe6 Qd7 23.Nxf8 Rxf8=+) 19...Bxd5! 20.Bxd5 e6! 21.Bc4 Qxd3 22.Bxd3 c2-/+. It must be noted that such performance during the game is, of course, beyond capabilities of a human being. 



Here Boris didn’t fail to comprehend that things have started to take a bad turn and made up his mind to change the course of the fight, and quite timely so: 


The 17...Nxd4 blow with the same ideas failed due to some very concrete reasons: 18.hxg6! Nxe5 (18...hxg6 19.Nxd4 Nxe5 20.Qh3) 19.Nxe5 Bxe5 20.gxf7+ Kg7 21.Rxh7+! Kxh7 22.Qh3+ Kg6 23.Qh6+ Kf5 24.Qh7+ Kg4 25.Qh3#. 

Passive defense, on the other hand, could have resulted in increasing of White’s mating possibilities: 17...e6?! 18.hxg6 hxg6 19.Nfg5 Qe7 20. Rh7!, and I will refrain from feeding you with kilometer-long lines of analysis that finally lead to +- or +/-. 


Apparently bad is 18.Nxe5? Bxe4 19.Qxe4 Qxd4-/+. 

18...Qxd3 19.Bxd3 a4! 

Now the regrouping of black pawns on the queenside starts to provide enough compensation for the sacrificed piece. It is also by no means unimportant that Black’s play is much easier. Black has a clear plan what he needs to do. White, on the other side, is in need of being on the constant lookout for defensive maneuvers or tactical possibilities to relieve his position through exchange of material, which is not always so easy.


This is a delicate moment. 20. bxa4! was a more practical continuation. 


In my opinion, White would be faced off with more practical problems in case of 20...axb3!?, for example, 21.gxf7+ Rxf7 22.Neg5 e6! 23.Nxf7 Bxf3! (rather than 23...Kxf7 24.Ng5+ Ke7 25.Rxh7+-) 24.gxf3 Kxf7 25.f4 Nd4. 

I would very much like to have a look at the hero who is capable of coming up with the 26.a4!! move in the practical game. This move was leading to a fantastic positional draw: 26...c5 27.Be3 b2 (27...Nc2+? 28.Bxc2 bxc2 29.Ke2 b3 30.Kd3 b2 31.Kxc2 bxa1Q 32.Rxa1+/-) 28.Rb1 Rxa4 29.Bxd4 cxd4 30.Rxh7 b3 31.Kd1, when neither side can achieve any progress to improve his position. 


This is a correct approach. 

21.Rb1? Rfd8 22.Bc2 axb3 23.axb3 Ra2–+ was going to lose the game. 21.Nc5?! would fail to resolve his problems in view of 21...axb3 22.Nxb3 Bd5 23.Bc2 Rfc8! to be followed up with с7-с7-с4. 


21...Rxa4 22.Nc5 Ra5 23.Nb3 Raa8 24.a3 Bd5 25.Bc2 would not be aggressive enough as further trade of pieces is going to insure White against any potential problems. 

22.Bc2 Bxa4 23.Bxa4 Rxa4 



This is a wrong approach. 

White’s king should have remained in the center in order to take part in the upcoming events: 24.Ke2! Ra5 25.Bf4 Nd4+ 26.Nxd4 Rxd4 27.Ke3, when the continuation 27...c5 28.Nxc5 Rxf4 29.Nb3 Rf5 that was being analyzed in the course of the online analysis is not as good as was initially believed due to 30.f4! and both black rooks find themselves being threatened. 

More prudent is 27...Rdd5, and then one of the possible continuations was 28.a3 g5! 29.Bxg5 (in case of 29.Bg3? c2 30.axb4 Rxa1 31.Rxa1 Rd1 it is the absence of the bishop on f4 that fails White) 29...Rxe5 30.axb4 Rxa1 31.Rxa1 f5 32.f3 fxe4 33.f4! – and White is going to escape with a draw in this ending. 


This is the only real inconsistency by Gelfand in this game. He let his b4-pawn stay unprotected for just a moment... 

24...Rd5! 25.Bf4 (25.Rb1 Bxe5) 25...c5 resulted in a serious advantage for Black. Keeping out all my other multiple lines of analysis I am going to provide below just one picturesque continuation: 26.a3 c2 27.Rfe1 (27.Rfc1 and 27.Rac1 did not escape our attention also) 27...b3 28.Nc3 Rxf4 29.Nxd5 Rxf3! 30.gxf3 Bxe5 31.Rxe5 Nh4! 

White is two rooks up but does not seem to be able to extricate himself from deep trouble! 

During the game, however, he succeeded in finding relief after 25.Rb1! c5 26.a3! 

So, White finally managed to break up the pawn chain of Black’s. The isolated pawns are much easier to cope with. 

26...bxa3 27.Nxc3 Nd4 28.Bf4 a2 29.Ra1 Rda8 30.Rfd1 g5 

30...f6 offered no winning chances either because of 31.Nxd4 cxd4 32.Ne4 fxe5 33.Bg5 etc. 

31.Nxg5 f6 32.Nge4 Nb3 33.Rxa2 Rxa2 34.Nxa2 Rxa2 35.exf6 Bxf6 

36.Rd7 Kf7 37.Rc7 Nd4 38.Bg3 Ne2+ 39.Kh2 Nxg3 40.fxg3 Bd4 and the game ended in a draw. 

Thumbs up for both opponents as it turned out to be a truly spectacular fight! 

So, Caruana, Dominguez and Svidler set the pace in the tournament after three rounds. The most exciting part of the tournament is still ahead of us...