Screwing with the structure: how to max out an initiative

A novelty deep down the road in a doubleedged, somewhat shady line of the Queen’s Indian, improving on a 2.700 GM’s game.

An instructive middlegame with a bunch of positional concepts and tactical motifs in play, pushing an initiative two pawns down.

And finally breaking the stiff resistance in an endgame, in which White sacrifices no less than four pawns in order to create a passer that makes it to the backrank eventually.

This one has it all. If I had to list my favourite games, this correspondence game from 2016 against a tough opponent would be on top.


Hero (2.222) – Villain (2.210)
Lechenicher Schachserver 2016, Queen’s Indian Defense

1. d4 Nf6 2. c4 e6 3. Nf3 b6


The Queen’s Indian Defense (QID), one of Black’s most reliable responses to 1.d4.

4. g3

White’s main move, but there are plenty more.


One of the ideas of 4…Ba6 is to lure the white queen away from the d file, then strike in the center with …c5, and White won’t be able to go d5 and lock Black’s light squared bishop out of the game.

5. Qc2 


The sharpest response, planning to go d5 nevertheless and sacrifice the pawn. 5.b3 would be the main move, followed by a mountain of theory. 5.Qc2 is known and played on every level for ages, but only had a brief period of increased popularity in the early 2000s.

5… c5

According to plan (I), the most principled move and the one that scores best according to the megabase. 5… Bb7 instead, intending 6…c5 (and not having to deal with d5) looks weird at first glance, but it makes sense.

6. d5


According to plan (II). With this sacrifice White aims for pressure on the d file and miscoordination in Black’s camp. Also after …exd5 the f5 square won’t be protected anymore. Especially a white knight enjoys the view on f5. If Black wants to cover this crucial square after …exd5, he will have to go …g6, creating further weaknesses (as can be seen in this game).

6…exd5 7. cxd5 Bb7

The immediate capture on d5 doesn’t work, it needs to be prepared.

(7… Nxd5?? 8. Qe4+)

8. Bg2 Nxd5 9. O-O Nc6 10. Rd1 Be7


The tabiya of the 5.Qc2 line. White has tried several approaches to crack the black camp, but over the course of decades all of them never lead to more than just enough compensation. 11.Qa4 became the main move in 2007 when Alexei Shirov introduced it in the Candidates tournament in his game against Levon Aronian, and the whole line picked up steam. In 2011 QID guru Sergei Karjakin finally defused it with his idea 14…Nd4! (see below) that leads to an equal endgame. Since then 5.Qc2 is a rare guest on the highest level once again, but there are continuous tries to revive the line. For instance the Russian supertalent Andrey Esipenko has made it part of his white repertoire.

11. Qf5

The old main move before people started to go 11.Qa4. Theory says, this isn’t much of anything for White either. But you know what, theory? I don’t care. It’s true, though, objectively 11.Qf5 barely gives White compensation for the pawn, nothing more. But it will lead to a long, complicated struggle with tons of possbilities for both sides, and there still is a lot to discover. Plus I had done some homework before entering this (sloppy homework, though, as you will see later).

(11. Qa4 Nf6 12. Nc3


(12. Nh4 O-O 13. Nf5 d5 14. Nc3 Nd4! = (diagram) Giving back the pawn, Karjakin’s move, introduced in 2011 in a game against Peter Leko (who later used it himself). This combined with 16…Qe8 is how Black equalizes against what had become one of the QID main lines in the mid 2000s. On first glance it looks like White may be able to develop some pressure against the isolated pawn on d5 but he can’t really. Objectively this is just equal, for instance 15. Nxd4 cxd4 16. Rxd4 Qe8 17.Qxe8 Rfxe8 18. Rd1 Bb4 19. Bd2 Bxc3 20. Bxc3 Rxe2, Nakamura, Hikaru – Harikrishna, Penteala, 1/2-1/2 (68 moves), Gibraltar Masters 13th, 2015)

12… O-O 13. g4 still pops up in super GM play occasionally, the last chapter in the 11.Qa4 try that the black players have yet to close. But since the big guys only play it in rapid or blitz events they don’t seem to trust White’s chances much. Funnily enough what today remains as the last open chapter in the 11.Qa4 line was also how its premiere game Shirov-Aronian went in 2007.)

11… Nf6 12. e4 g6


Black doesn’t love to weaken the black squares like this, but in the long run he won’t be able to avoid playing …g6 anyways.

13. Qf4 O-O 14. e5 Nh5


White’s guys find tasks quite easily – except the queen. Where to put it and what to do with it is one of the main topics of the 11.Qf5 line. 15.Qg4 is a decent alternative to what I played.

15. Qc4 Qc7

(15… d5 may be a decent try for Black that I failed to thoroughly check before entering this line. 16. exd6 Bxd6 17. Nc3 Na5 18. Qd3 Be7 probably doesn’t lead anywhere for White. He can try other queen moves on move 18, though. However, my opponent decided to go for something else.)

16. Bh6 Na5


We’re still in the realms of theory, and once more the question is: where to put the queen? But the constant queen moving doesn’t represent a loss of time. Note how both black knights have misplaced themselves by kicking the white lady around.

17. Qe2

(17. Qc2 is fine as well.)

17… Rfe8 18. Nc3 Ng7 19. Rac1


Already with 20.b4!N in mind.

(19. Nb5 2.700 Elo GM Igor Kovalenko probably is the strongest OTB player in correspondence chess. In 2014 he went 19.Nb5 and then pawngrabbing on d7 which didn’t lead to anything: 19… Qd8 20. Qd3 a6 21. Nc3 b5 22. Nd5 Ne6 23. Nxe7+ Qxe7 24. Qxd7 Qxd7 25. Rxd7 Rad8 = Kovalenko, Igor – Astroukh, Vitalij, 1/2-1/2 (30 moves), LSS 2014)

19… Bc6

(19… Ne6 20. b4 and 19… Rac8 20. Nb5 give White a substantial initiative.)

20. b4!


Novelty, not a hidden one unfortunately since this is a common motif. Longterm in these structures Black often manages to neutralize White’s play by installing a piece on d4 (a knight typically, see Nakamura-Harikrishna above). Taking this square away from him and opening the c file nets White an advantage even when two pawns down. Black just lacks coordination and active ideas. Stockfish doesn’t get it, though. For the machine 20.b4! is one possible move among many.

(20. Nd5 has been played before in correspondence chess.)

20… cxb4 21. Nd5 Qb7 22. Nh4!


The neccessary, somewhat counterintuitive follow-up that secures the white advantage. One purpose of 22.Nh4 is to fix the Nd5. Also it prevents Black from getting relief via …Nf5. White is ready to accept huge structural damage after …Bxh4 gxh4. In return he will get a lot of play against Black’s weakened dark squares.

22… Bxh4 23. gxh4


23… Nh5

Covering f6. Once again, note where the poor black knights are hanging out.

(23… Nf5 24. Nf6+ Kh8 25. Rxd7 winning)

24. Qb2


This move is less about regaining the pawn on b4 (which doesn’t run away) and more about dark square domination. The first mating patterns appear on the board, and they’re quite real already.

24… Re6

Not a nice square for the rook, but since no other blockers were around this is a sad neccessity. Otherwise White crushes by going 25.e6!.

(24… Rac8? 25. e6 Rxe6 (25… fxe6 26. Nf4 winning) 26. Nf4 f6 27. Nxe6 Bxg2 28. Rxc8+ Qxc8 29. Kxg2 dxe6 30. Qxb4 +-)

25. Qxb4


Once again this is less about capturing the pawn and more about keeping Black from getting the Na5 back into play. Black was planning …b5 and …Nc4. Now he can’t go …b5 since the Na5 would be hanging.

25… Rae8

(25… Rxe5 26. Ne7+ Kh8 27. Qd4 f6 28. Rxc6 +- Backrank mate!)

(25… Ng7 26. Nf6+ Kh8 27. Rxd7! winning)

26. f4


White is fully coordinated and brings further troops into the attack. Next on his to do list: lift the c rook to the f file. Black on the other hand has trouble to find a constructive move. When looking at this I wonder how Black managed to last another 30 moves.

26… a6 27. Rc2 Qa7 28. Rf2 Bxd5 29. Bxd5 Nc6 30. Bxe6


Cashing in since I didn’t find anything good to further up the pressure.

30… dxe6

(30… Nxb4? 31. Rxd7)

31. Qd6


This, however, is not a trivial win. White is now suffering from the grave structural concessions he made by allowing …Bxh4 gxh4. Now there are no good breaks on the kingside. Black will try to install a knight on d5 and set up some kind of fortress. If White wants to win this he needs to create play on both wings in order to have his rooks shine vs the more restricted knights.

31… Qa8 32. Rc1 Ne7 33. Qd7 Ng7 34. Bg5


(34. Rc7 Ngf5 35. Bg5 Rd8 36. Ra7 Rxd7 37. Rxa8+ Kg7 38. Bxe7 Nxe7 39. Rxa6 Nd5 (diagram) would lead to one of the fortresses I mentioned. White needs to play extremely carefully in order not to run into something like this that may turn out to be unwinnable (not 100% sure about this concrete one, though, but you get the idea). And White needs to be careful when it comes to trusting the engine. The machine doesn’t understand fortresses. As a general idea White would love to keep his bishop on the board in order to have it help organize play against Black’s queenside. And, of course, he must never give up his a pawn. Once there are only kingside pawns left, this is a draw immediately.)

34… Nef5 35. Rc7 Rf8 36. Qc6 Ne3 37. Qxa8 Rxa8 38. Rc6



(38… Nd5 39. Rfc2 Nf5 40. Rc8+ Rxc8 41. Rxc8+ Kg7 should be winning)

39. Rxb6 Kg7 40. Rb7 Rc8 41. Rd2 Nd5 42. h3 h6 43. Bf6+


Now the bishop may go, White even spits a pawn on top of it since he has a concrete winning setup in mind.

43… Nxf6 44. exf6+ Kxf6 45. Rdd7 Rf8


This is the position I was aiming for. I had analyzed for days in order to come as close as possible to being sure this is a win. White is ready to sacrifice his whole kingside in order to create a passer on the queenside that will have to win him the game. A delicate concept, but it seems to work out.

46. a4

The human has done plenty of work in this one, and successfully so. Time to have the engine finish Black off.

46… Nxh4 47. a5 g5 48. fxg5+ hxg5 49. Ra7 Nf3+ 50. Kg2 Ne5 51. Rd1 Rc8 52. Rf1+ Kg7 53. Rxa6 Rc2+ 54. Rf2 Rc3 55. Ra2 Nd3 56. Rb6 Nf4+ 57. Kf2 Rxh3 58. Ra4 e5 59. Rb1 e4 60. Rxe4 Ra3 61. Rb5 Kg6 62. Ke1 Ra2 63. Kd1 Nd3 64. Rd4 Nf4 65. Rc4 Nd3 66. Rc3 Nf4 67. Rc6+ Kh5 68. Rc4 Ne2 69. Rbb4 g4 70. a6 f5 71. Rc2 Rxa6 72. Kxe2


And since this is a tablebase position



Kommentar verfassen

Trage deine Daten unten ein oder klicke ein Icon um dich einzuloggen:

Du kommentierst mit Deinem Abmelden /  Ändern )

Google+ Foto

Du kommentierst mit Deinem Google+-Konto. Abmelden /  Ändern )


Du kommentierst mit Deinem Twitter-Konto. Abmelden /  Ändern )


Du kommentierst mit Deinem Facebook-Konto. Abmelden /  Ändern )


Verbinde mit %s