In round 10 of 14 in the 2013 candidates tournament, Magnus Carlsen played against Boris Gelfand. The winner of the tournament would earn the right to challenge the current world champion at the time, Anand.
The game was amazing. Some viewers even commented that Carlsen is better than the top computers engines. Humans must admit that nowadays it is almost impossible to beat the top computer engines at normal time controls.
But…
There is one thing that can’t be denied. In a critical position – which the top chess engines declared as being about equal – Magnus Carlsen played a combination of forced moves which gave him a solid advantage. The surprise is that the top computers couldn’t find these powerful moves as quickly as Carlsen did!
Let’s have a look at some of the critical and exciting stages of the game. The position below shows the critical position reached after Gelfand played 20… Qb6.
In this position, the top computer engines gave the position as almost equal. However, after a series of forced moves, even the computer evaluation had to admit that Carlsen achieved a solid advantage by using a series of powerful forcing moves. This proved that even the best computer engines can in some cases not yet match the positional understanding of the very top human players in the world.
The moves that followed serve as a great example of the philosophy to relentlessly give your opponent problems to solve:
Carlsen played 21.Bd4. Attacking the queen–move that essentially forces the black queen to go to b3, because all other options are worse.
Note that 21… Qxb2? (instead of Qb3) would be punished with the nice discovered attack 22.Nd5! White’s threats would be 23.Bxb2 as well as 23.Nxe7+. Black would not be able to solve both problems.
Note how Carlsen keeps on improving the activity of his pieces whilst at the same time forcing black to deal with threats. This is also called the initiative–a very powerful technique in chess.
I find it interesting that Gelfand played all the right moves till now and according to the engines, the position is still equal. However, the engines are wrong about this.
The only way to try delay white’s advance on the queen-side, is to play axb4.
Even though the engines recommend this move too, the engine evaluations are starting to favour white slightly. In other words, the engines now start to see why black may be in trouble.
White is again threatening b4-b5. On every move black has to deal with white’s threats! Since black doesn’t have a good way to deal with white’s threat, he goes for a counter-attack.
The downside to this counter-attacking move (which is again the engine’s top choice) is that the Nh5 is hanging (undefended.) Carlsen uses this consequence to keep up the threats.
Black needs to find a good defense to try prove that 24… Nh5 wasn’t just a pointless move.
However, the following exchanges are forced for black. Carlsen keeps it going.
This was black’s idea – Bxd4, Rxd4, Qxc3.
After the forced moves Rxd4, Qxc3 we reach this position:
Carlsen had yet another forcing move in mind that keeps black on the back foot.
Black must defend against white’s threats of Qxa8 as well as Rd8+.
Material is still equal but white’s pieces are much more active. The b7-pawn will be a target if Carlsen plays b4-b5, dislodging the bishop on c6. The point is that white took a seemingly equal position and by making threats on nearly every more managed to obtain a positional advantage. In this position the engines also agree Carlsen earned a decent advantage. It is worth mentioning that Gelfand played the engine’s top-choice moves throughout, but even so, the engines eventually agreed that this series of powerful, attacking and forced moves resulted in an advantage to white.
Carlsen skillfully continued to work with his advantage and score the full point by focusing his attention on the queen-side pawn majority. (If you’re interested to see the rest of the game, I included the pgn notation below).
[Event “World Championship Candidates 2013”]
[Date “2013.03.27”]
[Round “10”]
[Result “1-0”]
[White “Carlsen”]
[Black “Gelfand”]
[ECO “B30”]
1.e4 c5 2.Nf3 Nc6 3.Bb5 e6 4.O-O Nge7 5.Re1 a6 6.Bf1 d5 7.exd5 Nxd5 8.d4 Nf6 9.Be3 cxd4 10.Nxd4 Bd7 11.c4 Nxd4 12.Bxd4 Bc6 13.Nc3 Be7 14.a3 a5 15.Qd3 O-O 16.Rad1 Qc7 17.Be5 Qb6 18.Qg3 Rfd8 19.Rxd8+ Qxd8 20.Rd1 Qb6 21.Bd4 Qb3 22.Rd3 Qc2 23.b4 axb4 24.axb4 Nh5 25.Qe5 Bf6 26.Qxh5 Bxd4 27.Rxd4 Qxc3 28.Qa5 Rf8 29.Qb6 e5 30.Rd1 g6 31.b5 Be4 32.Qf6 h5 33.h4 Bf5 34.Rd5 Qc1 35.Qxe5 Be6 36.Rd4 Ra8 37.Qe2 Kh7 38.Rd1 Qc3 39.Qe4 Ra1 40.Rxa1 Qxa1 41.c5 Qc3 42.Qxb7 Qe1 43.b6 Bc4 44.Qf3 Qxf1+ 45.Kh2 Qb1 46.b7 Qb5 47.c6 Bd5 48.Qg3 1-0
A last note:
I am not trying to prove that humans play better chess than modern chess engines. It is widely accepted that it is nowadays almost impossible to beat the top computers at normal time controls. However, I believe this game proves that a deep understanding of chess is a critical skill which all top players must acquire – and at this time in history, the very top players can still, in some cases, find better moves than their computer counterparts.