GM Igor Kovalenko, rated with an Elo of 2.684 when this game was played, is the strongest OTB player in correspondence. He uses correspondence chess to test and sharpen his opening repertoire. Since results are secondary to him, he often doesn’t squeeze the last bit out of slightly superior positions (as long as he maintains a rating above 2.200 that allows him to compete in the highest league, I assume). Nevertheless, all of his correspondence games (the ones I’ve seen at least) are highly theoretical, many of them quite exciting.
This was my first game against him. How to approach it?
Usually I am happy to steer my games towards strategically complex, unbalanced structures that the engine has trouble to evaluate in order to maximize the human factor. But against a guy who completely outclasses me in every respect of the game I needed to change this. The general idea was to stay away from anything too closed, too strategical and get something rich and deeply tactical instead. The willingness to put more hours of analysis into it and catch him being lazy might be the only chance I have, I figured.
Already at move four I missed the chance to school a 2.700 grandmaster. I could’ve steered the game into a line he uses regularly in which I had found an almost crushing novelty a few weeks earlier that couldn’t have made it into his database yet.
Instead, I got schooled. This is the critical position at move 16 that we both had been aiming for:
If you want to take some training effect for your chess from this game make yourself familiar with it before you read any of the comments. First look at it for a minute, find the right move and assess the position, then look for another few minutes and assess again. It’s worth it, I promise.
Usually I restrict myself to posting correspondence games without too many of the usual superdeep tactics. This one is different. I put as many general remarks in there as possible and kept lines to a neccessary minimum. But that means that in some cases you will just have to trust my evaluation (or check it for yourself) since I didn’t want to get the analysis out of hand.
I’ve put all the moves and lines below into a Lichess study for you to replay and analyze more conveniently.
Hero (2.219) – Kovalenko, Igor (2.206)
Lechenicher Schachserver 2016, Queen’s Indian 4.a3
1. d4 Nf6 2. c4 e6 3. Nf3 b6 4. a3
The Petrosian System, named after former World Champion Tigran Petrosian, preventing …Bb4, preparing Nc3 and e2-e4, not my main approach against the ever so solid Queen’s Indian. But since I was confronted with 4.a3 from the black side in another game that had just started, I felt I could spare myself some work and get a better understanding of the system by playing it with White simultaneously. However, I soon became angry with this choice because I had forgotten about a recent Kovalenko correspondence game that I had just improved on big time.
(4. g3 Ba6 5. Qc2 is what I usually play. So does Kovalenko, who plays this with both colours. In one of his recent games he didn’t get anywhere with White in the exact same line where I have found an almost winning novelty that couldn’t have made it into his database yet. I should at least have tried to get him to play this line with Black and have him run into my improvement. This is how his recent game went: 5… c5 6. d5 exd5 7. cxd5 Bb7 8. Bg2 Nxd5 9. O-O Be7 10. Rd1 Nc6 11. Qf5 Nf6 12. e4 g6 13. Qf4 O-O 14. e5 Nh5 15. Qc4 Qc7 16. Bh6 Na5 17. Qe2 Rfe8 18. Nc3 Ng7 19. Nb5?! Qd8 20. Qd3 a6 21. Nc3 b5 and Black was fine in Kovalenko, Igor – Astroukh, Vitalij, LSS 2014. And now check what happened after 19. Rac1 Bc6 20. b4!N (diagram) in an earlier correspondence game of mine 😉 )
4… Bb7 5. Nc3 d5 6. cxd5
Sergej Karjakin likes to play 6…exd5, a move that is somewhat inflexible, dubious even in the eyes of most top players. Usually after 6…exd5 Black will have to play a hanging pawns structure. In the Candidates tournament 2016 the assembled world elite tried to prove that Karjakin’s love for this structure is misguided. They failed, and Sergej won the right to challenge Magnus for the title.
6… Nxd5 7. Qc2 Nxc3 8. bxc3
The starting position of the 4.a3 Queen’s Indian. White has invested two tempi to make moves that don’t quite fit the position: a3 and Qc2. He’d prefer the pawn on a2 and the queen on d1 (in order to move it to e2 later, then start a kingside attack). As compensation for the weird position of his a pawn and the queen White will get a beautiful e4/d4 centre and often use it to initiate an attack against the short castled black king. Back in the day Garry Kasparov has won many impressive attacking games in the 4.a3 Queen’s Indian.
8… c5 9. e4 Nc6
More common is 9…Nd7. 9…Nc6 is the modern, more direct move.
This felt like an easy choice, made by excluding the other possibilities. In the other 4.a3 game I was confronted with 10.Be3, and White didn’t seem to get anything at all. 10.Bd3 doesn’t seem to do much, so 10.Bb2 was the move left. When playing it, I was aiming for the position at move 16 where Black is forced to go 16…Kd8. That position, while extremely sharp, looked to me like I should be able to find something that exploits Black’s king position.
(10. Bd3 Rc8?? (better is 10… cxd4 11. cxd4 Rc8 with unclear play) In games between two world class grandmasters you rarely see someone being (almost) lost after ten moves. Even less so by overlooking a well known motif. 10…Rc8 is terrible, Black needs to open the c file first, only then go …Rc8. How to crush the immediate … Rc8 in this and similar positions shouldn’t be news to any Queen’s Indian player. 11. d5! (diagram) This one you can hammer out automatically. Black, a highly regarded theoretician, is toast. 11… exd5 12. exd5 Qxd5 13. O-O Be7 14. Rd1 Rd8 15. Bg5! +- Jakovenko, D – Bologan, V, 1-0 (21 moves), 17th Karpov Poikovsky 2016.)
Wait, what? Haven’t I just emphasized how important it is to go …cxd4 first, only then …Rc8? Yes, but in this concrete position White hasn’t developed the Bf1 yet and can’t castle immediately after opening the e file via d5 exd5 exd5. Still, the typical break 11.d5 is possible, but it doesn’t lead to an advantage.
(11. d5 exd5 12. exd5?! (12. Rd1 is better, leading to unclear play.) 12… Qxd5 13. Rd1 Qe6+ 14. Be2 Rd8 15. O-O Rxd1 16. Rxd1 Be7 and Black is better.)
11… cxd4 12. cxd4
Black can’t exploit the vis-a-vis of the Rc8 and the Qc2 yet, he needs to develop first, and quickly so. d4-d5 is a threat.
Only move. Everything else almost loses already.
(12… Be7? 13. d5 exd5 14. Bxg7 Rg8 15. exd5 Rxg7 16. dxc6 +-)
13. d5 exd5 14. exd5
(14. Bxg7? Rg8 15. exd5 Qe7+ and Black is much better. The option …Qe7+ is why he needed to play 12…Bd6, not 12…Be7.)
If White wants an advantage, this has to be the move.
(15. Qe2 White chickens out, this is just equal. 15… Qxe2+ 16. Bxe2 Ne7 17. Bxg7 Rg8 18. Bb5+ Kd8 19. Be5 Rc5 20. Bxd6 Rxb5 21. O-O Rxd5 22. Bg3 f6 = Susla, Vitaliy – Kovalenko, Igor, 1/2-1/2, LSS 2016.)
15… Na5 16. Qa4+ Kd8
(16… Kf8? 17. Qg4 is huge for White.)
It feels good to force the black king to stay in the center when 16…Kf8 looks like the more desirable move. White has reached the position he was aiming for. Now I started to seriously look for ways to crack the black position. On the first day of looking I didn’t find any, neither on the second and third.
The longer I looked into this the more it dawned to me that it was actually my opponent who had reached the position he was aiming for. His king on d8 is safe; it uses the white d pawn as a shield that protects him well. The white king on the other hand is not, and should White grab the pawn g7 it may quicky feel quite uncomfortable. It took me some days to realize that I had been way too optimistic when assessing this. White is balancing on the edge already, and he must fight hard to keep the balance.
Probably the most intensely analyzed move I have ever made – and a concession. White admits he has misevaluated things, changes plans, refrains from each of the many active possibilities and starts to fight for an equal or unclear game.
(17. Bxg7 Ba6! 18. Bf6! (18. O-O Bxe2 19. Bxh8 Rc4 20. Qb5 Bxf3 and Black is much better.) 18… Qxf6 19. Bxa6 Rc3 may hold as well, but it is clearly Black who is pressing. Again, the Kd8 is safe, the Ke1 feels under fire, wherever it goes.)
(17. Bc1 Ba6 18. Be3 Bxe2 19. Kxe2 Qd7 20. Qh4+ (20. Qxd7+ Kxd7 =+) 20… Kc7 21. Nd4 Kb7 =+)
(17. Qg4 =+)
(17. Nd4 leads to a clear black advantage.)
This had been played two times already, once in an engine game, once in a correspondence game. Both times White played the natural looking 18.Nd4?!, missing a devilish sequence of black moves that can easily be overlooked even with an engine running (see below).
Novelty. White happily exchanges g7 vs d5. Next on his to do list is to constantly threaten the h7 pawn. If White wants to stay in the game, he must turn his h pawn into a passer (and the Rh1 into a force instead of a bystander). In case the white a pawn falls, White needs quick counterplay on the kingside in order not end up in a lost endgame where the black a and b passers run too quickly.
(18. Nd4 Qe8! 19. Bb5 Ba6! 20. Nc6+ Qxc6! 21. Qh4+ Kc7 22. dxc6 Rxb5 23. Kg1 Rxb2 and Black is much better.)
18… Rg8 19. Bd4 Rxd5 20. Qc2 Re8 21. h4
Harry will have to run if White wants to stay in the game.
(21… Nc6?? 22. Bxb6+ winning)
(22. Bd3!? Ba6! 23. Rh3 Nc6 24. Bf6+ Kd7 (24… Kc7 25. Nd4 unclear) 25. Ne1 (diagram) may lead to an unclear game, but all the lines connected to it are so messy and complicated, I didn’t completely trust my judgement, even after analyzing it for quite some time. In the end I chickened out and decided to take the safe route towards a draw. My judgement had already failed me one time in this game afterall. In hindsight I think this is the route I should’ve taken since it fits the original gameplan: make it as messy as possible, then catch him being lazy.)
22… Rxe4 23. Rh3
(23… Bxa3 24. Rg3 harassing the f7 and h7 pawn gives White sufficient counterplay.)
24. Bc3 Rxd1+ 25. Bxd1
(25… Bxa3?! 26. Ng5 and White is much better.)
(26. Ng5 was possible as well.)
(26… Bxa3 27. Bc2 Ne3+! Only move that looks like the beginning of a nice combination, but White holds. 28. fxe3 Rc4 29. Bxh7 Rxc3 30. h6 Bc5 31. Nd4 Bxd4 32. exd4 Rxh3 33. gxh3 = It doesn’t look like it at first glance, but this ending is a dead draw.)
27. a4 Bc8
(28. Rh4 was tempting and probably good enough. But the h rook, even while passive at the moment on h1, still has a lot of potential. It is just waiting for the light pieces to capture the h6 pawn and will then help push White’s passer towards promotion.)
White has managed to keep it unclear. There is still an interesting, unbalanced endgame ahead, but Black offered a draw, I accepted. Just by the look of it it seems that Black has an edge, but White’s tactical possibilities against Black’s h pawn secure him equal chances.