Carlsen’s Clever Bad Bishop

Every chess position has pros and cons to it; it is up to the chess player to make the best of the situation and sometimes that means addressing the less favorable elements of your setup. Illustrating this concept, let us look at a game from the 2013 World Championship Candidates. This game was played on March, 14th with Boris Gelfand playing the White pieces and Magnus Carlsen playing the Black pieces. This game was played in round 3 of the event and we will follow Carlsen’s play as the less favorable “Bad Bishop” is addressed.

1. d4 Nf6 2. c4 e6 3. Nf3 d5 4. Nc3 Nbd7 5. Bg5 c6 6. e3 Qa5 This Queen move to a5 is a key element of the Cambridge Springs opening. In this opening, Black chooses to play for a little more control than the many sharp lines of the Slav Defense or Semi-Slav Defense and also avoids the common Queen’s Gambit structures that many 1. d4 2. c4 White players are accustomed to facing. This opening choice by Black is solid and many tactical tricks and traps exist for the unprepared opponent.

If given the chance, Black will usually continue with moves like …Bb4 and …Ne4 to put a lot of pressure on the pinned c3 Knight. There are also a few variations where Black captures on c4 and then the a5 Queen picks up a hanging piece on g5. Combating these ideas, White must play actively if they want a fighting chance.

7. cxd5 Nxd5 This variation is the Yugoslav of the Cambridge Springs. White’s most common continuations are by far either 8. Rc1 to reinforce the c3 Knight or 8. Qd2 to unpin the c3 Knight.

It is fairly simple to notice that other moves not addressing the c3 Knight fail. 8. Bd3?! Nxc3 9. Qxc3 has White dropping a pawn and 8. Bd3?! Nxc3 9. Qd2?? is not setting up a discovery on the Black Queen; it is the trapper who becomes the trap-ie after 9…Bb4! and White is lost.

8. Rc1 Nxc3 Carlsen opts for the 8. Rc1 line this time.

9. bxc3 Ba3 10. Rc2 b6 This …b6 move has the c8 Bishop in mind. It is instructional to usually develop as many pieces as possible. First develop the minor pieces, castle early, “connect the Rooks” and try to utilize the entire chess army at your disposal.

11. Bd3 Ba6! That is it! This is “Carlsen’s Clever Bad Bishop” moving to a6. The goal is to exchange the Bishop for White’s more active counterpart on d3. What makes this a “Bad Bishop” is that fact that Black has more pawns on light-squares now and this hinders the possibilities for the light-squared Bishop. This isn’t a terrible “Bad Bishop” because some positions feature far more pawns fixed on one color complex, but let us go back one move and assess what makes trading this Bishop on c8 such a good plan.

The e6 pawn is especially annoying to the c8 Bishop and this pawn is difficult to safely advance to the e5 square. Doing so would require a lot of preparation or else White will win material or liquidate some of Black’s potential influence in the center. Central control is very important in chess! If this plan isn’t feasible, then where else can the Bishop go? It's future on b7 looks questionable because it forces the fianchetto motif into a position that doesn’t call for it. We must play the needs of the position and not whatever elements we might feel like imposing at a whim. If the Bishop plans to become active on b7 then it would also require …c6-c5, but this doesn’t look that promising long-term. It is a slower idea and the a3 and a5 Black pieces appear inactive here.

It then makes sense to exchange off the Bishop with few prospects for the d3 counterpart that might become very active. It is all about piece activity!

We can’t give all of the credit to Carlsen though; this idea was not invented by him. In fact, this principle of exchanging inactive pieces for more active opponent counterparts goes back literally centuries. In fact, just in the decade from 2010 to 2019, my database counts six games where both players were rated over 2000 rating and reached this identical chess position after 11…Ba6!

12. O-O Bxd3 13. Qxd3 O-O Actually, all six of those games continued with these exact moves 12. O-O Bxd3 13. Qxd3 O-O in that order! It isn’t super surprising from a logical perspective though. The idea of …Ba6 is to exchange the inactive Bishop and both sides just capture and instructionally castle to complete development.

14. e4 Rfe8 15. e5 h6 16. Bh4 c5 Now Black plays to undermine White’s pawn center. Capturing the c5 pawn is positionally bad because it breaks up the pawns and leaves the e5 pawn weak, but not capturing allows …cxd4 and cxd4 to open the c-file (as soon happened in the game).

17. Nd2 cxd4 18. cxd4 Rac8 The chess game is still about equal for either side to win, but Black is making progress by chipping away at White’s central control.

19. Nc4 Qb5 20. f4 Rc7 The f4 move makes the White castled King a bit more exposed, but it does try to replace some central support that was chipped away at earlier. 20. Qb3 Qxb3 21. axb3 Bb4 seems like another way to play, but the game appears a lot more draw-ish with the Queens off of the chess board.

21. Qxa3 Rxc4 22. Rxc4 Qxc4 Now that a few more exchanges have taken place, this is a good time to re-asses the position. Why? Simple. The game state has changed and that means the plans of each side may or may not change, so it is worth a look at the very least.

This new position looks like both sides should place emphasis on relocating their pieces and especially so with the minor pieces. For White, that means the Bishop and for Black that means the Knight.

23. Qxa7? would be a mistake because it doesn’t “win” a pawn. Black responds with 23…Qxd4+ and regains the material, but what is more important: a central pawn on d4 or a backwards pawn on the Rook-file? Obviously the d4 pawn is more valuable! This means that White should do best to safeguard that central pawn and that is exactly what White does as they relocate the Bishop.

23. Bf2 Qc7 Of course, Black now defends their a7 pawn as they can no longer counter by safely taking on d4.

24. Rc1 Qb7 25. Qd6 Nf8 26. g3 Rc8 27. Rxc8 Qxc8 The game state must be roughly equal, but it feels like Black is making slight progress as they have now successfully taken control of the c-file from the White pieces. Now we shall see the other side make a small improvement to their position…

28. d5!? Of course! This interesting move gets rid of the backwards d-pawn and Black taking it also creates potential for the e6 pawn to become a dangerous passed pawn.

exd5 29. Qxd5 g6 Now that the endgame is here, both sides will usually try to activate their Kings if it is safe enough to advance them; ideally towards the center of the chess board where they can move to where the action might be.

30. Kg2 Ne6 31. Qf3 Kg7 Now the endgame is about to get fairly dangerous as both Queens are still on the board and both sides try to escort some pawns to promotion.

32. a3 h5 33. h4 Qc2 34. Qb7 Qa4 35. Qf3 b5 36. f5 gxf5 37. Qxf5 Qxa3 Perhaps the thrilling nature of this ending gives more justification for those “small improvements” because when the fireworks begin, it is sometimes the smallest of factors that prove decisive: things like a slightly safer King or a pawn protected that can save the game or win a game.

38. Qxh5 a5 39. Qg4+ Kf8 40. h5?! Qc1! Instructional. Perhaps GM Boris Gelfand (playing the White pieces) was rushed into playing 40. h5?! to meet the 40-move time control, but this slight inaccuracy is all GM Magnus Carlsen (playing the Black pieces) needs to gain the upper hand.

The reasoning behind 40…Qc1! is to instructionally prevent the advancing pawn from reaching the h6 square. Maybe a better try than 40. h5?! would have been 40. Qh5 when the ending still seems like a draw.

41. Qe4 b4 42. Be3 Qc7 43. Qa8+ Kg7 44. h6+ Kh7 45. Qe4+ Kg8 46. Qa8+ Qd8 Finally, the irksome checks are blocked and the Black King is nearby enough to the h6 pawn that it may be able to blockade it.

47. Qxd8+ Nxd8 I can’t really blame them for this because anything is simply losing, but it might have been more accurate to play 47. Qc6 and keep the Queens on the board for as long as possible. At least then, there is more hope for finding some perpetual check drawing resource.

48. Kf3 a4 49. Ke4 Nc6 The White King is doing their best and advancing to the center of the chess board as is common in many endgames, but it isn’t quite enough here because the King and Bishop will have a losing battle in trying to hold the connected passed pawns on a4 and b4 as well as the active Knight. The Black King is tied down to the guard of the h6 passer, but this is no trouble as Black’s overall forces should be decisive if they work together.

This is something we see from all good chess players: opening, middlegame, or endgame - they typically do a good job at coordinating their forces and this is especially prominent for the endgame when fewer forces are left, so there is more reason for them to work together.

50. Bc1 Na5 51. Bd2 b3! The a5 Knight is immune to capture because then the Bishop can’t return to stop the pawns in time. For example: 52. Bxa5? b2 53. Bc1 b1=Q+ 54. Ke3 a3 and it is clear which army is better in this endgame.

52. Kd3 Nc4 Once again, the Knight is immune to capture because …b2 and promotes.

53. Bc3 a3 54. h6 Kh7 55. g5 Kg6 Both pawns on h6 and g5 are blocked by the King.

56. Bd4 b2 57. Kc2 Nd2 0-1 White resigns. Again, the Knight is immune to capture and now the Knight also controls the crucial promotion square b1. Yes, White could sacrifice the Bishop on b2 for both pawns, but then the f7 pawn will win the game as the d2 Knight can collect all the enemy stragglers while the g6 King holds the line.

A possible continuation to checkmate might go something like:

58. Bxb2 axb2 59. Kxb2 Nc4+ 60. Kc3 Nxe5 61. Kd4 Nf3+ 62. Ke3 Nxg5 63. Kf4 Kxh6
When Black simply needs to slowly escort the pawn to the f1 square:

64. Ke3 Kg6 65. Kf4 Ne6+ 66. Kf3 Kf5 67. Kf2 Kf4 68. Kg2 f5 69. Kf2 (but the defending King taking the opposition is not enough to draw because Black has Knight moves available; otherwise, this variation would be a draw - therefore, Black would play this ending differently without the Knight) Nd4 70. Kg2 Ke5 with forced checkmate in 17 just before …Ke5 is played according to Stockfish if both sides play perfectly.

71. Kg3 f4+ 72. Kg2 Ke4 73. Kf2 f3 74. Kf1 Ke3 75. Ke1 f2+ 76. Kf1 Ne6 77. Kg2 Ke5 78. Kg3 f1=Q 79. Kg4 Qf3+ 80. Kh4 Qg2 81. Kh5 Qg5#

The long endgame had enough steam to convert a win out of the small advantage gained after 40. h5?! Qc1! but it was Carlsen’s Clever Bad Bishop move 11…Ba6! which resonated with me. It nicely highlights the ever-changing environment of the chess board and the constant need to seek piece activity. This usually means improving and coordinating your own forces, but occasionally (as we saw here) it means exchanging off your inactive pieces for active counterparts from the enemy.

It might prove to be just enough to snatch the game like what happened in this case. The smallest of advantage and one side may press for the full point.