For those of you who were formally introduced to chess like me, you may recall being taught the importance of the solidarity in pawn structures. The more fragmented a structure becomes, the more pawn islands are created. Since pawns are “stronger” together, it’s logical then to believe that each pawn island (or isolated pawn) created thus weakens the integrity of one side’s overall structure. This static consideration is so important that many coaches for beginners say that the side with fewer pawn islands can be considered better! While this grossly undervalues the power of dynamic play, this consideration can help steer the structurally better player in the right direction.
In the case of endgames, understanding this principle is crucial, as a brittle structure offers various targets throughout the duration of the game. In our previous Endgame Essentials posts, we discussed how a weak king or a badly placed piece can single-handedly change a result. By simultaneously asking yourself how you can improve your position and stop the opponent’s counterplay, we can try to stretch out (or limit!) our opponent’s defensive resources by creating a passed pawn, or dominating an opponent’s piece. When taking structures into consideration, often times we don’t need to immediately create our own attacking resources because they are already provided for us. As we have with our past studies, we resume our travel through Magnus Carlsen’s career – resuming in 2009, and today reaching the year 2011.
As we move through each exercise, I encourage you to continue asking yourself how Carlsen can improve his position. When playing against a weak structure, the duration of the plan will take longer, and usually a win is not simply obtained by tactical means like some of our previous examples.
Carlsen – Karjakin, 2009
At a first glance, neither sides’ pieces are particularly impressive. Karjakin’s rook on d8 seems to stand strong on the d-file, but as we’ll see in a second, it actually has no entry square on the d-file that’s particularly useful. To get a better assessment of who’s better, we move to the theme of today’s lesson by comparing structures. In the purest definition of the word, each side has exactly three pawn islands. However, the value of each island is different. For example, visually, we can already see how the isolated c6 pawn is a lot weaker than White’s on h3. By being on a half-open file, Black’s c-pawn can present him with immediate problems. Furthermore, I think something needs to be said of Black’s e5 pawn. While at a basic level it belongs to the same pawn island as the f-, g-, and h- pawns, supporting it with another pawn would actually be a concession for Karjakin. Already, the pawn on e5 limits the scope of Black’s dark-squared bishop. Should Black ever play …f7-f6, he limits the bishop even more, while White’s opposite colored bishop improves.
So as we can see, while Carlsen also has three pawn islands, it doesn’t limit his ability to improve his position. 21. Nd1 Rd6 22. Rc5 Kf8 23. Kf1 h5 24. Ne3 Ke7 25. Ke2 Bg7
Both sides have tried to improve the position, but White’s done a better job of addressing Black’s weaknesses. From c5, Carlsen’s rook hits both the c6 and e5 pawns. Without a clear improvement, White spends this move asking himself “what’s my worst piece?” and finds that the knight on e3 has limited mobility despite its centralization. With 26. Nc2 Carlsen makes a move he’ll have to make anyway to reactivate the knight while waiting on Karjakin to find improvements 26…Bh6 27. Ra5!
Why not immediately take the pawn on e5? Carlsen decided here that given the choice, he’d rather win the pawn on a7. Should White win this pawn, not only does he get a passed pawn on the a-file, but the pawn on e5 still blocks in Black’s bishop. Karjakin didn’t let this happen, but protecting the a-pawn means retreating one of his pieces. Carlsen wasn’t worried about 27…Rd2+ 28. Kf1 Rd1+ 29. Kg2 and with no more checks, Black must go back and protect a7. It’s in this line that we see how Black’s rook isn’t really a factor on the d-file.
27…Rd7 28. Rxe5+ Kd6 29. Ra5 Bg7 30. f4!
Giving Karjakin a choice. By taking the pawn on b2 like he did in the game, Black temporarily puts his bishop offside and has to spend several tempi reactivating it. Meanwhile, White can still put pressure on c6 and a7. While Karjakin’s chances for survival dwindle by playing the role of materialist, he doesn’t exactly have a better option.
30…Bxb2 31. e5+ Ke7 32. Nb4 Kf8 +=
In ditching his c6 pawn, we can safely say that Carlsen holds an advantage. Had Black tried to hold on with 32…Rc7? 33.Rxa7! Rxa7 34. Nxc6 still gives White a nice two pawn cushion. White doesn’t even have to be flashy because 33. Rc5 will win on c6 as well – if 33…Kd7 34. Bxf7 +-.
33. Nxc6 Bc1 34. Kf3 Rc7 35. Rc5 Ba3 36. Rc2
After spending the last few moves to regroup, Carlsen’s ready to move onto phase two of this endgame. While White stands a pawn up, given the nature of rook and minor piece endgames, there’s still more work to do. The most immediate solution is to try to find ways to make the e-pawn passed. With White’s bishop on b3, it’s important to keep an eye out for sacrifices on f7, but there’s time to improve the position first. Since Black lacks any light square control, White can play to isolate Black’s f7 pawn with Kf3-e4, and f4-f5 with an edge. While this never happened in the game, I’m sure Carlsen saw it (the engine approves too!).
36…Nc8 37. Ke4 Kg7 38. Bxf7!
Though the idea of 38. f5 would have won slowly, this move immediately points out Black’s lack of coordination. Karjakin must take back on f7, and whichever way he chooses, he allows Nc6-d8 with a discovered attack on c8. Even with two minor pieces for the rook, Black doesn’t have enough to slow White’s passed pawn.
38…Kxf7 39. Nd8+ Ke8 40. Rxc7 Kxd8 41. Rc3 Bb4 42. Rd3+ +-
And now for phase three – creating more passed pawns. By trading the f4 and g6 pawns, Carlsen can have connected passed pawns, thanks to his other f-pawn on f2. Once this happens, Magnus will push the e- and f-pawns until Black’s minor pieces stop immediate advances. The remainder of the game is added for the sake of completion.
42…Ke7 43. f5 gxf5+ 44. Kxf5 a5 45. f4 Nb6 46. Rg3 Nd5 47. a3 Be1 48. Rd3 Nc3 49. e6 a4 50. Rd7+ Ke8 51. Rd4 Ke7 52. Ke5 Nb5 53. Rxa4
Sure, White has a passed a-pawn now too, but if anything, this is just a confirmation that Black has lost.
53…Bc3+ 54. Kd5 Nc7+ 55. Kc4 Bf6 56. Ra7 Kd6 57. f5 Ne8 58. Rd7+ Ke5 59. Rd5+ Ke4 60. a4 Nc7 61. Rd7 Ne8 62. e7 1-0
Black must now give up a minor piece to stop White’s passed pawns, after which White’s rook and a-pawn will prove enough.
This endgame was particularly instructive because it shows the uncomfortable decision Black must constantly make between material and activity. Here Karjakin was consistently compliant with Carlsen’s pawn grabbing, but once the position opened, White was able to use his passed pawn (like our earlier endgames) to limit Black’s play and win. In our next game, Carlsen faces Ivanchuk in a rook and knight endgame where the Ukranian was adamant to hold onto his material.
So again we have a position where piece play is relatively even. Each sides’ rooks are planning to contend for the c-file and are arguably worth the same at the moment. While White’s knight seems menacing on d4, it can only move backward. Black’s knights have a similar issue as it’s unclear as to where they belong. If we do a basic pawn island count, we can see that Carlsen has two, while Ivanchuk has three. So where in the position is White’s structural advantage giving Carlsen an edge? The d4 square. Since Black’s d5 pawn is isolated, that means a pawn can never kick a piece from d4. However, we already mentioned that the knight here doesn’t offer much for White. When our opponent’s pawn structure doesn’t give us enough to work with, the next step is to see if we can create new targets. This is why Carlsen played 39. h5! and after Ne7 40. Rh1 gxh5 41. gxh5, we’ve reached a new structure.
Even though White’s created an isolated pawn of his own, Black now has three isolani in the position. I think it’s interesting to note how the engine still considers this endgame equal. Perhaps in a perfect world this position is tenable, but in practice this isn’t so easy to hold – and that should be enough for White. Carlsen’s plan is to activate his rook via h1-h4-f4 to attack f7, and then push his queenside pawns to create another weakness.
41…Rg8 42. Ng3 Rg5 43. b4 Kd7 44. Rh4 Ne8 45. Rf4 Nd6 46. a4 b6
Black creates a padlock here and has done well thus far to improve his position. Black’s rook is a little awkward on g5, but it’s doing a good job of pressuring White’s only concession as a result of the structure change seven moves ago. Meanwhile, the knight on d6 offers Black mobility, with ideas of …Nd6-c4, putting pressure on e3, making sure the king stands guard.
47. a5 bxa5
This is more or less forced, as 47…b5 48. Nb3! with the idea of reaching c5 and pressuring a6. By trading on a5, Ivanchuk eliminates this permanent outpost.
48. bxa5 f5? +=
Black’s woes begin here with this committal move. Already it was becoming difficult to find improving moves for White, so simply waiting with 48…Re5= would have forced Carlsen to come up with new ideas. The Ukranian’s move is a mistake because it moves his weakness within reach of White’s knights, making it easier for Carlsen’s pieces to create pressure. I’m thinking Ivanchuk just panicked here because Rf4-f6 can be met with …Ne7-g8 and Black holds.
White’s rook is no longer needed on f4 since White’s knights are watching Black’s f-pawn. By activating the rook White can play to infiltrate on the queenside. Black can bring his rook over too, but that means no pressure on h5, and fewer defenders of the f5 pawn. Before relocating the rook, Carlsen will insert f3-f4 to stop any potential pawn sacrifice ideas of …f5-f4 and fix the weakness.
49…Nc4 50. f4 Rg4 51. Rh3 Nd6 52. Rh1 Rg8 53. Rb1 Ra8
While it may not seem like much has happened, all of Black’s pieces are tied to pawns, giving White time to do the one thing he’s done best: improve his pieces.
54. Kf3 Kc7 55. Ne6+ Kc8 56. Nc5 Rb8 57. Rxb8+ Kxb8 58. Nxa6+
A good rule of thumb for knight endgames is that often times they can be calculated to a result like pawn endings. While this can be impractical to do over the board, being up a pawn in a knight endgame is definitely a promising sign, and in this game, Carlsen manages to convert. For the sake of brevity, I want to skip to a critical moment.
58…Kb7 59. Nb4 Nc4 60. a6+ Kb6 61. Ke2 Nd6 62. Kd3 Nb5 63. Ne2 Ka5 64. Nc3 Nc7 65. Nbxd5!
Sacrificing the knight! Thanks to the spread of White’s pawns, Black is not in time to stop promotion. Being able to sacrifice the knight to simplify into a won endgame is an important resource, and it’s definitely not an uncommon endgame idea. The game continued:
65…Nexd5 66. Nxd5 Nxd5 67. a7 Nc7 68. Kd4 Kb6 69. Ke5 Kxa7 70. Kxf5 Nd5 71. Kg6 Nxe3 72. Kxh6 1-0
Black’s king is too far to stop White’s pawns, so Ivanchuk resigned here. Unlike the Karjakin game, Ivanchuk held onto his weaknesses (and rightfully so!), only to err later with 48…f5?. In retrospect it seems like a simple mistake, I think it’s really illustrative of how difficult it is to play such a position and just hold.
In today’s post, we discussed how a simplistic understanding of pawn islands can help us find weaknesses and weak squares. Similar to having better pieces, having a better structure can give you control of the pace of the game, ultimately making the difference between a win and a draw.