So far in my Endgame Essentials series, I’ve laid out some basic principles to improve our overall assessment of different positions. Understanding that our opponent has a weak king, sidelined piece, or a cancerous structure can help us seize the initiative and identify a plan going forward. While the examples I’ve previously given are relatively straightforward, in practice, such applications are not so simple. Take this position from the recent Candidates Tournament for instance:
Svidler – Karjakin, 2016
In the game, Svidler made the logical move, 48. Rxf4, after which the game followed 48… Rxa2 49. Rfh4 g6 50. Re5 with a draw. I don’t think it’s fair to compare White’s choice to that of an engine, but Stockfish’s recommendation here is particularly instructive – 48. Re5! with a big plus for White. The point is that after 48…Rxa2 49. Re7 g6 50. Rxf4, White’s rooks are a lot more active than Black’s and now both the e3 pawn and the 7th rank are weak. Furthermore, Black’s knight on f8 is out of commission with no pleasant square for refuge. Again, it’s hard to fault Svidler for the miss, but the engine shows us here that activity is stronger than material (for more of my thoughts on engines, here’s a post from last year).
What this should tell us is that the heuristics we’ve identified thus far should always be at the forefront of our attention. However, sometimes we don’t have the convenience of having a better position. In such cases, one strategy is to strengthen our structure by gaining space in the aims to restrict our opponent. If I had to choose a “one-move” example of this, it would be from this past year’s Tata Steel.
Navara – Caruana, 2016
At a first glance, the position is seemingly equal. Navara has a broken pawn structure, but his activity offers enough compensation. If White had moved the bishop here from d5, Black would immediately take the second rank with …Rd6-d2!, seizing the initiative and potentially the game. This is why Navara chose 35. c4!, protecting the bishop, but also showing Caruana how inactive his rook really is. From d6, the Black rook has limited options, and can’t easily put itself on the e-file. The Czech player went on to win a very nice endgame, and I encourage you to see its continuation here.
Naturally, improving a pawn structure takes more than one move, but I thought this case illustrates the aims of the expanding side quite nicely. As we have throughout this series, we’ll take a look at a few examples from Magnus Carlsen’s past victories, this time from 2012 and 2013.
Carlsen – Van Wely, 2013
Already, we have a messy position. White has the bishop pair, but the light-squared bishop seems a little boxed in on d3. The most glaring weakness in this position is the f5 pawn, but Loek has set a trap: 23. Bxf5 Ne5!= and despite being down a pawn, the constant pressure on c4 is enough to give Black equality. But as I hope you’ve noticed thus far, the endgame rewards long-term plans more than short calculations, so this pawn on f5 will be a source of concern for Black going forward. Just remember, sometimes the threat is stronger than the execution! So Magnus instead chose 23. Kc2 (Though imprecise, 23. f4 should win too since it covers the e5 square) Bd4 24. Rb1 Nb6 25. Bf4
Before deciding on a structure, Carlsen has decided to optimize his pieces. By putting pressure on his opponent first, he will have a better idea of what structures will give him the best winning chances.
25…Be5 26. Re1 Kg7 27. Bg3!
The starting point for today! With this move, Carlsen intends f3-f4, fixing the weakness on f5, and limiting his opponent’s bishop’s mobility. Already, holding the file and keeping his position intact is getting uncomfortable.
27…Re7 28. f4 Bf6 29. Rxe7 Bxe7 30. Be1
Relocating the bishop the long diagonal is a clear idea, but Carlsen wants to gain space on the kingside with his h- and g-pawns. Again, there’s no rush to take on f5, the pawn can’t go anywhere, thanks to the pawn on f4.
30…h5 31. g3 Bf6 32. Kb3 Kg6 33. h3 1-0
Perhaps it was premature, but Van Wely resigned here in light of 34. g4, finally winning the f5 pawn. With the bishop pair and a healthy material advantage, White should win with relative ease.
This is an important endgame because it shows us that long-term weaknesses can usually not be held by tactical means forever. White maximized a static advantage by fixing the f5 pawn and trading rooks, making it difficult for Black to create counterplay.
In our next example, Carlsen takes on Caruana in a position that is much more balanced:
Carlsen – Caruana, 2012
In this position, both sides have exactly one weakness. For White, the isolated c-pawn is a clear target, and for Black, the backward pawn on b6 is also an issue. I think here many players would try to exchange weaknesses, but, in this case, this mutually beneficial trade will only result in equality (Note that the immediate 29. Bxb6 fails anyways to 29… Rxe1!, I mean this as a more long-term idea). But here it could be argued that White’s position is simpler to play. The bishop on d4 is better placed than it’s counterpart on c7, and can’t easily be kicked from its outpost, thanks to the c3 pawn. Furthermore, it’s much easier for Magnus to put pressure on b6 than it is for Caruana to attack c3, so Black still needs to prove equality in this position. Knowing this, White decided that it was time to expand on the kingside.
29. Re4 g6 30. g4 Kf8 31. h4
Even though it’s not yet clear how Magnus will use these pawns, we can say that he has improved his position, and now asks Black how he will relieve pressure on the b6 pawn. Caruana starts with an exchange and quickly claiming the e-file.
31…Rxe4 32. Kxe4 Re8+ 33. Kd3 Re6 34. Be3!
And now it’s starting to become clear how Carlsen intends to use his kingside pawns. Should Black push ahead with 34…h5?! 35. gxh5 gxh5 36. Rb5 +=, White can enjoy a long-term advantage with pressure on both b6 and h5.
34…Kg7 35. Rb5 Bd8 36. h5
Even though Caruana has made completely natural moves, White has consistently made matters difficult for him. Should Black try 35…f4, he will constantly have to defend a weak h6 pawn. Meanwhile, White can change gears and play c3-c4-c5, only now trading weaknesses because it will be more difficult to defend a5 and h6 than it currently is with b6 and h6. Black decided to keep his structure compact, but this means his king is stuck on g7 protecting h6 until the structure is resolved!
35…Rd6+ 37. Kc4 Rc6+ 38. Kd5 Re6 39. Bd4+ Kf8 40. f4 +=
Black has some weaknesses, but nothing nearly as pronounced as our previously analyzed games. However, by improving his pieces and getting space on the kingside, White’s advantage is already becoming visual. Black now is challenged to find moves that don’t make concessions.
40…Bc7 41. f5!
Pressuring the g6 pawn. White’s intention is to make the h6 pawn much more exposed. Even if Caruana tries 41… gxf5 42. gxf5 with the belief that White’s structure also becomes weak, he’ll quickly find that he has no easy way of attacking the isolated f- and h-pawns, since b6 (and soon h6) are under fire. Sometimes, your opponent’s biggest weakness is only as weak as your strongest strength – here the damage to White’s structure is negligible.
41…Rd6+ 42. Ke4 Rc6 43. Rb1 Ke8 44. hxg6
Now that Black has distanced himself from his kingside pawns, Carlsen takes on g6 with the h-pawn so he can attack h6 via h1.
44…fxg6 45. Rh1 Kf7 46. Kd5 Rd6+ 47. Kc4 gxf5 48. gxf5 Bd8 49. f6!!
An incredible interference! White trades the kingside pawns, with the idea that liquidating pieces will only help White since his king is closer to the queenside. Black has to oblige, and as we’ll see, his position quickly collapses.
49…Bxf6 50. Rxh6 Be7 51. Rxd6 Bxd6 52. Kb5
And 23 moves later, the debate is resolved, the b6 pawn was weaker than the isolated c-pawn. It was important that White expanded on the kingside because it came with the caveat of having a better king in the final position. Black played on for another 14 moves, but the win is simple. Carlsen picked up the last of Black’s pawns and then pushed his down the board.
For our last example today, both sides attempt to expand in the endgame, but Carlsen’s opponent tried for too much – which ultimately proved for his own demise!
Carlsen – Svidler, 2013
Already, it’s move 12, and we have a queenless middlegame. Black’s bishop looks a little silly on g7, but other than that, we have relative equality in the position. If Black were on the clock, Svidler would likely choose …Bc8-e6 limiting White’s e2 bishop, so Carlsen started with 12. Bc4. Svidler, needing to get his c8 bishop into the game with 12…b5 (which engine thinks is fine), but based on the game’s continuation, Black already puts himself in a place where he must be extremely accurate. White doesn’t really have any threats, which is why I prefer 12…Bd7, with the idea of rerouting to c6. It takes just as many moves as Svidler to develop, just without the bonus of a forcing move. One of the reasons I don’t like this move is because of a general principle Grandmaster Magesh Panchanathan once taught me – don’t move pawns for short term plans. It’s not clear yet if this queenside expansion is beneficial to Black, and as we’ll see Carlsen successfully punishes him later. 13. Bb3
Already we can see some reasons as to why 12…b5 may be questionable. First, b3 isn’t exactly a “worse” square than c4 for White’s bishop. More importantly, the move a2-a4 is beckoning to be played, with the idea of undermining Black’s structure.
13…Bb7 14. f3 Bf8 15. a4!
Now Svidler is faced with an uncomfortable decision. Does he take on a4 and cripple his queenside forever, or does he hyperextend with b5-b4? While the b-pawn push is optically pleasant, it comes with the drawback that c4 is weakened forever.
15…b4 16. Nb1
Taking advantage of Black’s hyperextension. Carlsen plans a quick maneuver, Nb1-d2-c4 to put pressure on e5.
16…Nd7 17. Nd2 Bc5 18. Kf2!
A nice application of a simple idea here – trade only if it helps you! Taking on c5 would activate Black’s knight, so now, if Svidler wants to trade dark-squared bishops, he must take on e3, activating the king!
18…a5 19. Rfd1 Kg7 20. Nc4 Bxe3+ 21. Kxe3 f6 22. Rd2
Phase 1 of White’s plan is complete. Magnus stands slightly better thanks to his control over c4, but Svidler has done well to not create new weaknesses. The next stage of the game is brief, as Carlsen simply grabs the d-file.
22…Nb6 23. Nxb6 cxb6 24. Rad1 Rxd2 25. Rxd2 Bc6 26. Rd6 Rc8
Once again Carlsen is doing well, but it still seems like Svidler can hold this position. In phase 3, White finally improves his structure on both sides of the board to increase his winning chances.
27. Be6 Rc7 28. b3 Kf8 29. Bc4
White has sealed the queenside, as now both a4 and c2 cannot easily be hit. Meanwhile, b6 is already a future target for White. But first, Carlsen plays on the whole board!
29…Kg7 30. h4 h5 31. g4?
Svidler must make another tough decision. Does he take on g4, allowing White the opportunity to create a passed h-pawn in the future? Or does he allow White to take on h5, creating another target? As it turns out, Black actually missed a chance to equalize here with 31…hxg4! 32. fxg4 Bxa4! 33. Rxb6 and Black has a lot fewer weaknesses in the position. Carlsen was better if he found the prophylactic 31. Bd3!, removing the idea of …Bxa4 and planning an f3-f4 push. The endgame is still complicated, but White still has an edge.
31…Bxa4 32. Rxb6 Bd7 33. gxh5 gxh5
While Black may have gotten rid of his b6 weakness, he now has targets on a5, f6, and h5. Even though Black isn’t lost here, White is still for choice.
34. Bd3 Kf7 35. f4 exf4+ 36. Kxf4
Winning this endgame won’t be simple, but by trading the e5 pawn for his f-pawn, Carlsen opens up dark squares in the center for his king. After getting his rook onto a better square, Magnus centralizes the king by moving it to d4.
36…Rc5 37. Rb7 Ke6 38. Ra7 Kd6 39. Ke3 Bg4 40. Kd4
Now with a centralized king, White has slightly better winning chances. White will now bring his bishop to c4, reducing the Black rook’s options.
40…Re5 41. Bc4 Bf3 42. Bd5! +=
With this move, White is in full control once again. Black’s rook only has two moves that don’t immediately drop a piece, and each of Svidler’s three weaknesses are much more difficult to defend.
42…Bd1 43. Rf7 Re8 44. Rxf6+ Kc7 45. Ra6 Bxc2?? 46. Rc6+ 1-0
A simple oversight by Svidler in a position that was already lost. This game gave us both good and bad examples of expanding the structure. Early in the game, Svidler pushed too quickly, giving White counterplay on the queenside and a great outpost on c4. But this wasn’t enough to win. By expanding on the kingside (the one blunder aside), Magnus managed to break Black’s pawn structure.
In today’s post, we discussed how in seemingly equal positions, we can increase our winning chances by improving our pawn structure and gaining space on each side of the board. Often times it isn’t enough to have one weakness in the position, so often changing the structure (in our favor) gives us more attacking options and plans to stretch out our opponent.
I’ll be playing my first tournament in over a month this weekend at the Marshall Chess Club in New York City, so I won’t be able to post my usual video on Sunday. Look out for my next post early next week, where I’ll hopefully be sharing what turned out to be a good performance!