[Event "World-ch Carlsen-Karjakin +1-1=10"] [White "Karjakin,Sergey"] [Black "Carlsen,Magnus"] [Site "New York"] [Round "9"] [Annotator "Caruana,Fabiano"] [Result "1/2-1/2"] [Date "2016.11.23"] [WhiteElo "2769"] [BlackElo "2853"] [PlyCount "148"] {The game after Carlsen's loss wasn't guaranteed to be exciting, but viewers were in for a surprise. I had expected a fairly tame draw, but instead we were treated to an extremely complex fighting game with an opening not usually seen at the top level.} 1. e4 e5 2. Nf3 Nc6 3. Bb5 a6 4. Ba4 Nf6 5. O-O b5 6. Bb3 Bc5 {The Archangel variation was certainly not expected by most people, and probably not by the Karjakin camp either. Playing it requires a tremendous amount of knowledge, so it's likely that Carlsen prepared it before the match as a backup to his main defenses. I have some experience playing this as Black against Karjakin, as it was my mainstay opening for a time, and we played several interesting games. As an attempt to win, it can work if the opponent is poorly prepared or surprised, but Karjakin proved up to the task.} 7. a4 Rb8 8. c3 d6 9. d4 Bb6 10. axb5 axb5 11. Na3 {Karjakin decided to go for the main and most principled line, immediately grabbing the b-pawn. It is what he recently played against Svidler and the most forcing variation, so he probably felt playing this way avoided the most amount of risk.} O-O 12. Nxb5 Bg4 13. Bc2 exd4 14. Nbxd4 Nxd4 15. cxd4 Bxf3 16. gxf3 Nh5 17. Kh1 {I always felt that this line was one of the most unpleasant for Black to face. White avoids any risk and hopes to prove the advantage of the bishop pair after the center opens up, usually when White sacrifices a pawn with e5 or f4. Meanwhile, Black has to constantly be on the lookout for tactics.} Qf6 18. Be3 c5 $5 {Years ago, as Rustam Kasimdzhanov was frying some plantains, he explained to me that this move was the best chance for Black to equalize. After we analyzed it, he tried it against Nakamura in the 2014 Tromso olympiad. Carlsen and his team also must have felt this was the best way for Black to approach the position.} 19. e5 Qe6 20. exd6 c4 $1 {This unusual move is the point behind 18...c5. Black keeps the center closed, and plans on capturing d6 next and creating mating threats after with ...Bc7.} 21. b3 cxb3 (21... c3 {was how Rustam approached the position, and although it may be playable, he lost the game. Carlsen's choice is perhaps a better way to equalize.} 22. d5 Qxd6 23. Ra6 Nf4 24. Ra4 Ng6 25. Qd3 Bc7 26. f4 {Nakamura-Kasimdzhanov, Tromso ol 2014, and Black was outplayed.} )22. Bxb3 Qxd6 23. Ra6 {After 23 moves, we have reached a critical position, and one which was certainly analyzed by both players and their teams. Magnus sank into thought here, which perhaps means that he forgot what he had prepared, or that he was deciding between two roughly equivalent options.} Rfd8 ({Another option was to play} 23... Qd7 {immediately, and considering the Bxf7 tactics which happened later in the game, keeping the rook on f8 for now was worthy of consideration.} 24. Rg1 g6 {I have a feeling that this would have offered Black better chances for equality.} )24. Rg1 Qd7 25. Rg4 {I thought that placing the rook on h4 was very strange, but there was no other convenient way of defending the d4-pawn. The next dozen or so moves are very difficult to explain. The position is highly concrete and it's likely both players were creating and reacting to immediate threats; the end result being very computeresque play.} ({The tempting sacrifice} 25. Rg5 g6 26. Rxh5 gxh5 { seems to lead nowhere.} )({The computer suggestion of} 25. Bc4 Bxd4 26. Bg5 { would lead to a very drawish ending:} Rdc8 27. Ra4 Rc7 28. Bxf7+ Qxf7 29. Rxd4 Rf8 30. Rd8 Rc8 31. Rxc8 Rxc8 {, and White's winning chances are minimal} )Nf6 26. Rh4 Qb5 27. Ra1 g6 {This move is generally useful, providing luft for Black's king.} 28. Rb1 Qd7 29. Qd3 Nd5 30. Rg1 {Now White's plans are clear. Rooks on the open files and a bishop on b3 pave the way for an eventual rook sacrifice on g6, h7 or a bishop sacrifice on f7. The immediate threat is Bg5.} Bc7 31. Bg5 Re8 32. Qc4 Rb5 $1 {An excellent move, but I'm not sure it was one which was accurately calculated by either player.} (32... Nb6 $4 33. Qxf7+ Qxf7 34. Bxf7+ Kxf7 35. Rxh7+ {shows the tactics available at White's disposal.} )33. Qc2 (33. Ba4 {is of course critical. White seems to win a rook, but Black has a sneaky response:} Qf5 $1 34. Qf1 $1 {Interposing with the queen on g2 is the only way to save White. Now Black has only one move to stay in the game:} Rb1 $3 (34... Qxf3+ 35. Qg2 Qxg2+ 36. Kxg2 {wins for White.} )35. Qxb1 Qxf3+ 36. Rg2 Nc3 37. Qf1 Nxa4 {, and Black has very decent compensation for the exchange in the form of White's uncoordinated forces and a mighty queen on f3. However, White could still try to regroup and play this for a win, even though a draw is the most likely result.} )Ra8 $6 {After this, Black again experiences difficulties.} (33... Rb4 {is the machine recommendation, the point being to keep the rook on the e-file for} 34. Bd2 Re2 $1 )34. Bc4 Rba5 35. Bd2 Ra4 36. Qd3 Ra1 37. Rxa1 Rxa1+ 38. Kg2 Ne7 $2 { Almost the decisive mistake, but practically this move was difficult to punish. } ({After something like} 38... Bd8 39. Re4 Nf6 {, Black is still clearly worse, but he has decent chances to hold on due to counterplay against White's king.} )39. Bxf7+ {Very tempting, but not correct. After this sacrifice everyone in the spectators' lounge got very excited and started predicting a win for Karjakin. However, the whole line becomes very forcing and leads to a drawish position.} (39. Qb3 $1 {was more critical:} Nf5 (39... Qf5 40. Re4 Bd6 41. Qb7 {just wins material - the knight is trapped.} )40. Bxf7+ Kg7 (40... Qxf7 41. Qxf7+ Kxf7 42. Rxh7+ Ke6 43. Rxc7 {, and this ending is most likely lost or at least very difficult to hold.} )41. Rh3 {It feels like Black should have counterplay here, but there is no path to create threats to White's king. It turns out Black's king is the terminally weak one.} Qe7 42. Bg8 h5 43. d5 { , and Bc3 is coming. Of course, this would not guarantee a win, as the position remains very complicated, especially in time trouble, but objectively Karjakin would have good winning chances like this.} )Kxf7 40. Qc4+ Kg7 41. d5 ({My first thought during the game was} 41. Rxh7+ Kxh7 42. Qf7+ Kh8 43. Bh6 {, with mate I thought, but then I saw} Qxd4 {and the tables are turned. Black wins.} )(41. Bh6+ Kf6 $1 {is surprisingly also bad for White, as the attack leads nowhere.} )Nf5 {Now everything is forced.} (41... Be5 $2 42. Bc3 Qd6 43. Qf4 $1 {is a very study-like win, with a double pin on Black's king and queen.} )42. Bc3+ Kf8 43. Bxa1 Nxh4+ 44. Qxh4 Qxd5 {And here it became clear that Carlsen would not lose the game. White is not winning h7 and his pawn advantage is not enough to cause Black significant problems. Karjakin tried hard to create progress here, but to no avail.} 45. Qf6+ Qf7 46. Qd4 Ke8 47. Qe4+ Qe7 48. Qd5 Bd8 49. Kf1 Qf7 50. Qe4+ Qe7 51. Be5 Qe6 52. Kg2 Be7 53. Qa8+ Kf7 54. Qh8 h5 55. Qg7+ Ke8 56. Bf4 Qf7 57. Qh8+ Qf8 58. Qd4 Qf5 59. Qc4 Kd7 60. Bd2 Qe6 61. Qa4+ Qc6 62. Qa7+ Qc7 63. Qa2 Qd6 64. Be3 Qe6 65. Qa7+ Ke8 66. Bc5 Bd8 67. h3 Qd5 68. Be3 Be7 69. Qb8+ Kf7 70. Qh8 Qe6 71. Bf4 Qf6 72. Qb8 Qe6 73. Qb7 Kg8 74. Qb5 Bf6 {Finally Karjakin gave up trying to win. Magnus survived a difficult position! If he had lost, the match would have been practically decided, but now it is still a wide open contest. Tomorrow Carlsen has a critical game with White to try and win; he hasn't managed to convert his advantages so far in the match, so we'll see if he can pull through in the crucial last games.} 1/2-1/2