FIDE World Cup 2017: Why didn't So sacrifice an exchange?

by Sagar Shah
9/20/2017 – Wesley So showed some excellent technique in the queenless middlegame to completely outplay his Chinese opponent Ding Liren. On the 40th move Ding made a mistake which gave Wesley an opportunity to sacrifice the exchange and torture his opponent for the rest of the game. Wesley had seen the move, yet did not play it. We try to find the reason why the American GM missed this opportunity. Aronian could not show anything concrete against MVL's Grunfeld and the players quickly agreed to a draw. Game analysis, pictures and exciting videos from Tbilisi. | Photo: Amruta Mokal

The Art of the Positional Exchange Sacrifice The Art of the Positional Exchange Sacrifice

The positional exchange sacrifice is one of the most powerful and fascinating strategic weapons in chess. On this DVD Sergey Tiviakov explains why the positional exchange sacrifice is such a strong weapon and how to use it.


Two draws at the Semi-finals

World Cup

The Hualing hotel is unusually empty. Only four players are left at the event. The number of journalists and media is also not very high. Six days ago, a playing hall that had 16 players, is now now down to just two tables. Although the number of people has diminished, the intensity and importance of every move has increased manifold. At the end of two days of classical chess and one day of tiebreaks (if required) we will have two players qualified for the Candidates 2018. Who will the be? According to the polls we had in our previous article, there is little distance between MVL and Aronian. But in the case of Wesley So and Ding Liren, our readers definitely think that the former is a big favourite.

The guard outside the playing hall doesn't recognize Wesley So!

So vs Ding

Wesley So and Ding Liren, one of these guys will play in the Candidates for the first time in their life in 2018! | Photo: Amruta Mokal

If people thought So was a favourite, they were right, at least from the way the first game progressed. It was a Guioco Piano where things were around equal after the first 15 moves. Wesley then found an interesting idea:


Black should have let the queen be. But instead Ding exchanged the queens and allowed White some clear plans to improve his position.


White's next plan was to get the rook from e1 to a1 and then shift his f3 knight from e1 to c2 and prepare the b4 break. Not feeling good about these moves, Ding went for ...a4. Now Ra1 can always be met with ...Nb3, but the a4 pawn surely becomes weak.


This was a very strong move, but also a very interesting decision. Usually when you make such important decisions you would like to have time on your clock. But So didn't make any waiting moves to reach the 40th move mark. He just played the best move in the position and was confident about his chances.


The error couldn't have come at a worse time for Ding. 40 moves had just been completed and So had 30 minutes on his clock. Of course, Wesley had seen the best move in the position, but did not execute it. He played Kc3 and Liren found all the accurate moves needed to ensure a draw. The question is, could White have sacrificed an exchange on b3 and played for a win? I ask this question in the video below:

Million dollar question: Was Wesley So winning or not?

Update: The million dollar question has been solved! The win has been found for White. But not with the move 43.Kc3, but 43.Rg8! Check the analysis below our interview with Ding.


When I went for dinner I let my engine run, and when I came back I was confident about the evaluation of most of the variations as you can see from the depth of the engine.


It is pretty easy to look at active moves like Rg7 for White here, or Nxf6. But what is not simple to see is 41.Kc3! The idea: don't let Black establish both his rooks on the seventh rank. However, during the game prophylactic moves often go unnoticed. Later when I told this move to Ding, he was unaware about it and was very surprised that it existed. But within a fraction of a second he realized its strength. Wesley in his interview with press officer Anastasiya Karlovich said that he had seen the move Kc3, but thought it was not so easy to win. Perhaps objectively 41.Kc3 might not be 100% winning, but it gives White a risk free position and a chance to play for only two results. An opportunity that was surely missed by So.

Ding Liren speaks to us after the game

I have tried to analyze the critical moments of the game in quite some depth and I now share my analysis with you:

Update, September 20th: ChessBase readers  especially Mark S helped in finding the truth of the position and 43.Rg8! instead of 43.Kc3 - a prophylactic move which I thought was very strong. Here's the updated analysis. 


A relieved Ding Liren after the game | Photo: Amruta Mokal

Ding Liren knows that he was very close to defeat. But now he has the white pieces and in the last two matches (against Wang Hao and Richard Rapport) he has been absolutely ruthless when he has had the first move. Wesley will have to tread carefully against his preparation.

Aronian vs MVL

Levon ties his shoe laces and gets ready for the marathon! | Photo: Amruta Mokal

I somehow have the feeling that MVL has good chances in the World Cup format against Levon Aronian. Two things work in MVL's favour: strong nerves and more energy. If there is one thing that Aronian has a clear edge over his opponent, it would be superior opening preparation. Levon usually has new ideas up his sleeve. But in the first game his opening preparation was pretty dismal. We will come to it.

Initial moments of game one from the semi-finals

Look at Aronian's clock! They have played 23 moves and he has seven minutes more than what they started off with!

Looking at the clock one might think that Aronian had prepared everything from the white side of the Grunfeld. As it turned out, there was hardly anything to worry for Black. In fact at some point it felt as if White simply had no chances. The two queenside pawns were just too strong and it wouldn't be inappropriate to call this highly unsuccessful home preparation by Aronian.

MVL had no difficulties in holding the balance and drew without any issues. After the game he said, "I had looked at this line earlier this morning. I just didn't understand what Levon was trying to achieve. I had this game against Nisipeanu where both of us played the same line, but with the inclusion of the move ...h6. So I am well versed with the theoretical updates in this line." 

MVL and Levon are good friends | Photo: Amruta Mokal

Levon has the unique style of making his move with one hand and writing with the other! How do you like his green coloured shirt! Photo: Amruta Mokal


Tactic Toolbox Grünfeld

The most effective, timeproven way to develop tactical abilities, imagination, and the ability to calculate variations, is practice. The 69 exercises on this DVD are taken from grandmaster games and show tactical ideas that are typical for the Grünfeld.


MVL's coach at the event Etienne Bacrot said that Maxime could have made his opponent's life tougher with some accurate play, but the evaluation (0.00) would not have changed much | Photo: Amruta Mokal

Maxime Vachier-Lagrave speaks about his game with Levon Aronian: "I just didn't understand what was going on!"

Expectations and Statistics

(click to enlarge)

38 games: 10 wins for MVL, 7 wins for Aronian and 21 draws (including one today)

In the past whenver Maxime has opened his game with 1.e4, Levon has replied with 1...e5. The players have played eight games in the Ruy Lopez. In their last encounter MVL went for 3...Bc4. This led to a win for the French grandmaster in the London Chess Classic. The chances of us seeing Giuoco Piano tomorrow is pretty high. I have a feeling that Aronian wil be able to hold it with the black pieces tomorrow against MVL. If the match does go into the tiebreaks the younger of the two — Maxime Vachier Lagrave will surely have an edge. 


(Click or tap to enlarge)

17 games: 2 wins for Wesley So, 2 wins for Ding Liren with 13 draws (including round one here)

It's very interesting to note that the first game that So and Ding played with each other was way back in 2004 at the World under-12 Championships in Heraklion. Ding won that encounter. The next time they sat opposite each other was in the first round of FIDE World Cup 2011 in Khanty Mansiysk. Both of them had a rating of around 2650. So managed to clinch that match as he won the first game of the rapid tiebreaks. In 2015 Wesley scored another win over Liren at the Bilbao Masters. In 2016 the pair played four game match in Shanghai. Ding was able to win that match 2½-1½. 

The Chinese player likes to open his games with 1.d4 and there are high chances that we might see Catalan in action tomorrow. Who do you think will win?


Sagar is an International Master from India with two GM norms. He is also a chartered accountant. He loves to cover chess tournaments, as that helps him understand and improve at the game he loves so much. He and is the co-founder and CEO of ChessBase India website, the biggest chess news outlet in the country.
Discussion and Feedback Join the public discussion or submit your feedback to the editors


Rules for reader comments


Not registered yet? Register

jasontaylor7 jasontaylor7 9/21/2017 05:01
I don't mean to be harsh but this article is mostly bs since as others have shown it fails to correct an error of the 41. Rxb3 line analysis that So failed to play but did consider. As I wrote on youtube: In this video at 2:45 Shah claims that the continuation of 41. Rxb3 would be
41. Rxb3! axb3
42. gxf6 gxe6
43. Rg7 ,

but this is actually wrong. That last move would be blunder even though it gets a pawn. A non-blundering continuation is

43. Rg8+

which moves black's king, allows pawn exchanges, and ultimately a win for white. Shah has even admitted to this on the sidebar of his commentary, but I guess he doesn't want to admit his mistakes more openly since it is hidden in a sidebar.
Kilovs 2016 Kilovs 2016 9/21/2017 03:56
Go Wesley!!
Raymond Labelle Raymond Labelle 9/21/2017 03:30
Very interesting observation Aighearach. Something to try maybe: make the moves and hit the clock with your non-dominant hand and write with your dominant hand. Levon can even make both things at the same time. We may not always need to save that much time in classical time control, but who knows?
hserusk hserusk 9/21/2017 10:21
It's a great example of having your cake and eat it too. The top players (especially the 2750 and above) are supposed to be chess gods and only tourneys against each other would be worth the sponsorship money. On the other hand when they miss a shot and a win, we're told that they are human too.
The world championship cycle meanwhile has reduced number of rounds than the great ones in the past we all know of and the last three have been a real drag. Still we're to understand that this is the absolute best we could ever do in seeing a top flight chess match.
Loving the doublethink. Aces!
Michael Jones Michael Jones 9/21/2017 12:06
You've pretty much answered your own question with the words "When I went for dinner I let my engine run". While letting engines spend hours analysing a position can be fascinating as a theoretical exercise, it has absolutely no relevance to practical over the board play. If it takes an engine minutes or even hours to find the win, a human, even one of the world elite, has almost no chance of finding it at the board. I guess that's why So didn't play it.
Raymond Labelle Raymond Labelle 9/20/2017 09:49
"Why didn't So sacrifice an exchange?" Because it was very hard to see and he did not benefit of the assistance of a computer.
Aighearach Aighearach 9/20/2017 08:43
Levon is using the most efficient method of moving the pieces and hitting the clock; the clock is to his left, and you're supposed to move the pieces and hit the clock with the same hand. There are probably a lot of us that move the pieces with the hand on the clock side of the board and yet who still always write the moves down using our dominant hand.
geraldsky geraldsky 9/20/2017 03:38
The blunder there was the the guard. He must be replaced before he'll make a trouble.
Sagar Shah Sagar Shah 9/20/2017 08:37
Mark S, thank you for the update. I had some sleep and woke up to this and have updated the article with the analysis. Rxc7! was especially nice.
Mark S Mark S 9/20/2017 07:19
Even better,
41. Rxb3 axb3 42. gxf6 gxf6 43. Rg8+! Kd7 44. Nb4! Ke7 45. Rc8 Ra2 46. Nxa2 Kd7 47. Rxc7+! Kxc7 48. Nb4 +-

The final position on this improved line would result to Knight versus Rook endgame where N will defeat the Rook because of extra pawns by white. +6 according to engines.
This position wins for the Knight.
TommyCB TommyCB 9/20/2017 07:04
When I was analysing with Stockfish, 43. Rg8+ immediately jumped to +3 or more. How did Komodo and Sagar Shah not see this? So much effort wasted analysing only 43. Rg7.
Mark S Mark S 9/20/2017 06:19
Someone posted in a forum that Rxb3 exchange sac of white would win with a +5 for white.
The line was: 41.Rxb3 axb3 42.gxf6 gxf6 43.Rg8+ Kd7 44.Nb4 Ke7 45.Rc8 Ra2 46.Nxa2 Kd7 47.Ra8 bxa2 48.Rxa2 Rh2+ 49.Kc3 Rh1 50.Ra8 Rc1+ 51.Kd2 Rg1 52.Kc2 Rg7 53.Kc3 +-

But might be not that easy to find over the board.
Raymond Labelle Raymond Labelle 9/20/2017 05:35
Not totally on subject, but, by coincidence, at the end of the guard incident video, there is a link to the final moments of the Magnus - Bu game which Bu won. Up to the departure of the players. Both players act correctly and in a civilised manner.

Wanted to specify it because I read here and there that Magnus did not look OK in these final moments- well, this is probably an unfounded rumour, as this was filmed. Here: