Game Six Carlsen Anand

Game 6

Game six  featured an incident - the double-blunder, which will be trotted out on a regular basis whenever this match is discussed in future. It also featured an odd opening choice by black and some insane defensive sidelines which only the engines saw. Eventually, the man who made the second-to-last error won.

Carlsen deserved to win. Make no mistake about that. The world champion had continuous pressure out of a cleverly-managed opening and he kept that edge into the queen-less middle game. The weirdness on move 26 was an aberration. Much worse errors have been made in world championship matches - people have missed forced mates, forced wins of pieces.  

But this error comes in the post-computer engine era and that means it will be discussed at great length by many people who would not see it, or calculate it through unaided.

Consider the Diagram after 26. Kd2?

Black "should" play 26. -- Nxe5!  Some calculation is necessary even after spotting that move.
Black needs to see the trivial  27 Rxe5? Rxg4 would win an exchange. That  ensures that the forced white response is 27. Rxg8 and then Nxc4 CHECK 28. Kd3 (attacks the Ktc4) Nb2 CHECK ( and the Kt cannot be attacked again) 29. King  moves somewhere Rxg8. This leaves black two pawns ahead. Further, black has to ensure that his Kt b2 is not trapped - it can exit to safety via a4/ c4.    

This 3-4 move deep sequence is not difficult to calculate . But it is not trivial either. Of course Anand would have seen it, and seen it quickly at that. if he had been looking for it. But then Carlsen would have seen it too if he had been looking for it.

So the question is, why were they not looking for it? One answer is that the position was very, very good for white and neither player thought that evaluation would change easily. In fact, even after 29.--- Rxg8, white is not dead-lost even two pawns down in an endgame.  White will get the h6 pawn back and his bishop pair could help him hold a draw. And that is a measure of how good white's position is, before the blunder.

Incidentally, both players forgot the "Blumenfeld Rule". Biniamin Blumenfeld (1884-1947) was a minor Soviet chess master. He invented the sideline known as the Blumenfeld Counter-Gambit (1.d4 Nf5 2. c4 e6 3. Nc3 c5 4. d5 b5). He also wrote a short, very influential essay,  " Visual Imagination and the Calculation of variations" which carried a lot of sound advice including the following rule.

The rule: When you have finished a long calculation,  before playing the first move, take a fresh look at the board through the "eyes of a patzer". Also, when your opponent makes a move, even one that is expected, you should never (unless in extreme time trouble) without thinking, immediately make the reply you have thought about.
In both cases, Check if you are missing something trivial. 

The blunder-counter-blunder obscures the fact that Anand made a mystifying opening choice. The Sicilian Defence is fine, the Paulsen is fine. A variation that exchanges queens on move 9 is not fine.  More so, since White actually has a serious advantage and black must either grovel or take serious risks to limit that edge.  

Anand has played this sort of "semi-equal" variation before, against  Topalov's queens pawn in the 2010 World Championship for instance.  (http://www.chessgames.com/perl/chessgame?gid=1581339) was a loss but two earlier tries with the same opening were draws.  The key difference is that Topalov doesn't really like slow positional grinds. Carlsen loves queenless middlegames of this type.

Carlsen leads 3.5-2.5 in the match and he has another white coming. Anand's first priority would be to recover his mental equilibrium - that is an entirely internal process. His second priority would be to find a less painful defence and here he needs to discuss what happened with his seconds.

This match is not over but this game was definitely a saddle-point.  

Some interesting moments

Diagram after 12. -- Kc7

White has space and the bishop pair, with  a near-bind on the dark squares. It's not clear how black can untangle his bishop. It is also not clear how black can defend the dark squares on his king-side. The doubled c3,c4 pawns are not much compensation for this.

Kramnik said this position was "Unplayable with black for a human being" after admitting he had looked at it. Svidler, (who also plays the Paulsen regularly) said that Anand would have had to study the following plans for white and he found it puzzling why black would go here.

The engines say white has a small advantage but the engines may be wrong. This is the kind of position where white will implement over-the-horizon plans that the engines don't see. Also, unlike humans,  engines don't have a psychological problem or energy issues with playing continuous defence. It would be interesting to Monte Carlo this position out with strong engines playing each other many times. The evaluation might change to "white has a large advantage".

Diagram after 14. h5. Black can block the h-pawn with h6 as he did in the game. After that, his g7,h6 pawns are targets. He did not make another appreciable error  until move 26-- but white just got steadily better.

The other possibility is to allow h6. White will not necessarily play h6 instantly (or he might). He might build pressure via Rh4-g4 and Rd3-g3. His control of g5 will also translate into a grip on d8 since Bg5 can be played at some stage to ht d8. That may make it impossible for black to even challenge the d-file if white doubles at some stage. It's a moot point which line (playing h6 or allowing white to play h6 at his choosing) gives black better chances of survival. He is under pressure whatever he does.

 We've already discussed the blunder. What about alternatives for white instead of 26. Kd2?

Diagram after 25-- Rdg8.

This is what Anatoly Karpov used to call a "50-50 position". Karpov would win 50 percent of the time with white and draw the other 50 per cent of the time.Carlsen might see it as 70-30 in favour of a white win. 

Moves like 26. a3 or 26. Rg3  just retain a large plus for white. Black is holding tactically if white plays 26. Rxh6? Nxe5 and Black hopes he can generate enough compensation if white plays 26. Bxh6?!  since the white pieces are a little tangled. 
 

Diagram after 28.-Rd8. Black's hoping that his threats of hitting a2 in an opposite-coloured bishop endgame will compensate for material losses on the kingside. He might have tried 28..Ra8!?.
There is a strange tactical trick which neither player saw, nor did the commentators (they weren't using engines of course).

After 28...Ra8 29.Bc1 Ba4 30.Be4+ Kc7! 31.Bxa8 Rxa8 32.Bxh6 Bc2 or 32.Bxa3 Bc2 33.Bc1 Bf5 34.Rg3, black seems to have serious drawing chances. Why? See the next note.

Diagram after 32. Be4+

A version of the same exchange sacrifice idea is playable with 32.--Ka7! 33. Bxa8 Kxa8 34.Bxa3 Rd1 35.Rxh6 Ra1. There is no clear way to extract the Ba3 and black has drawing chances since 36. Ke3 Nxe5 37. Rg7 Nxc4+ complicates. Another possible continuation is  36.Rg5 Rxa2+ 37. Ke1 Rxa3 38. Rh7 Rxc3 39. Rxf7 Bc2 40. Rf6 Rxc4 41. Rxe6 Kb7. This will boil down into a pawn race and it's not so clear who wins that. Hurrah for the engines!  This was really the last chance.

( The game score for reference )

Carlsen  - Anand  [B41]

Game 6 WCM Sochi 2014, 15.11.2014

1.e4 c5 2.Nf3 e6 3.d4 cxd4 4.Nxd4 a6 5.c4 Nf6 6.Nc3 Bb4 7.Qd3 Nc6 8.Nxc6 dxc6 9.Qxd8+ Kxd8 10.e5 Nd7 11.Bf4 Bxc3+ 12.bxc3 Kc7 13.h4 b6 14.h5 h6 15.0–0–0 Bb7 16.Rd3 c5 17.Rg3 Rag8 18.Bd3 Nf8 19.Be3 g6 20.hxg6 Nxg6 21.Rh5 Bc6 22.Bc2 Kb7 23.Rg4 a5 24.Bd1 Rd8 25.Bc2 Rdg8 26.Kd2 a4 27.Ke2 a3 28.f3 Rd8 29.Ke1 Rd7 30.Bc1 Ra8 31.Ke2 Ba4 32.Be4+ Bc6

33.Bxg6 fxg6 34.Rxg6 Ba4 35.Rxe6 Rd1 36.Bxa3 Ra1 37.Ke3 Bc2 38.Re7+ (1–0).