(Another) Mid-Week Special Edition: Nakamura defeats MVL to Win the Chess.com 2020 Speed Chess Championship
Of note: Carlsen's rivals are rounding into form...
I wrote last week about Nakamura-So in the Semifinal of the Chess.com 2020 Speed Chess Championship, and I intended to cover the other semi-final and/or the final of that event this week. But there’s also the Russian Superfinal going on - where Nepomniatchi (who sits tied with MVL atop the currently-suspended candidates tournament cross-table) jumped out to a clear lead by defeating former World Championship challenger Sergey Karjakin in Round 7. And there’s the Russian Women’s Superfinal going on, where Shuvalova and Goryachkina are tied for the lead ahead of a crowded field.
So! (ha, chess pun) we are going to cover the remaining rounds of the Speed Chess Championship (“SCC”) in a special, mid-week edition, and free up the weekend’s column for a serious look at some serious games from the Russian national championships.
Team Solomid’s Hikaru Nakamura sits atop the world, at least, online, for now.
Last week we left our coverage with Nakamura’s declaration - his match against Magnus would be on Saturday at noon. Doh! Magnus was ousted from the tournament in the semi-final round by one Maxime Vachier-Lagrave, ranked #5 in the world (according to live rankings, where Nepo has passed MVL due to his performance in the Russian Superfinal - more on that this weekend) in classical time controls, #2 in rapid, and #3 in blitz. “MVL” as he is colloquially known, will also hope to close out his early success in the candidates tournament and earn the right to challenge Magnus in an officially sanctioned World Championship match. But this weekend, we got MVL vs. Magnus Carlsen in a marathon match of blitz and bullet games. MVL proved that, should he challenge Carlsen for the world championship, the champ might not want this match to go to rapid or blitz tie-breaks…
Let’s see how it went down:
Match of the (mid)-Week #1:
Maxime Vachier-Lagrave vs. Magnus Carlsen (Chess.com 2020 Speed Chess Championship Semifinal)
MVL is a bit unique amongst top level players in his loyalty to a rather narrow opening repertoire. He nearly always plays 1. e4 with the white pieces, and he nearly always plays the Grunfeld with black against 1. d4, or the Sicilian Najdorf against 1. e4. One of the great intrigues of a potential MVL-Carlsen world championship match, actually would be: does MVL dare to play the openings that he’ll be certain Carlsen has prepared for? Or does he venture into “unfamiliar” territory to try to create some type of advantage out of his home preparation? The world may get to find out, or it might not. In the first game of the 2020 SCC semifinal Carlsen had the white pieces, played 1. d4, and we got a Grunfeld. After white’s 13th move, MVL had a big decision to make:
This position will be familiar to Grunfeld players, at least in the abstract. While Carlsen played a rather rare line with an early Qa4+, the positional characteristics of the position are familiar. White has a big space advantage in the center due to the push d5!?, which has forced black to misplace his c6 knight. The poor piece has moved already 4 times in the first 13 moves (Nc6-b4-a6-c5), and still sits loose on the c5 square. For his troubles, black has already castled, the long diagonal is open for the g7 bishop, and he’s maybe one or two moves away from completing his development. Black would like to (1) find a permanent home for the loose c5 knight (2) develop the light squared bishop - probably to g4, where it will be exchanged for white’s knight and (3) undermine white’s imposing, but advanced central pawns.
The problem is that these things cannot be done “in any order.” Precision is required from MVL. The computer likes a materialistic solution that doesn’t feel very natural to me, as a Grunfeld player myself. By giving up the prized g7 bishop, black can basically win the e4 pawn, and hold onto it until white exchanges his dark squared bishop for the c5 knight. But this feels very dangerous - if white isn’t forced to exchange the dark squared bishop for the knight, and simply ignores the extra material, the dark squares around black’s king could become very very weak… Stockfish and Komodo agree, though:
… Bxc3!? 14. bxc3 f5! 15. c4 (15. exf5 Bxf5 16. Qd1 Qxd5) fxe4 16. Nd2 Qd6 17. Be3 Bf5!? (holding onto the pawn for now) - this is likely to be black’s best continuation.
But even the “human” 13. … f5!? should provide equality:
… f5!? 14. e5 (again, 14. exf5 Bxf5 16. Qd1 Ne4!, and white is having real problems along the long diagonal) Ne4 15. Nxe4 fxe4 16. Qxe4 Bf5 17. Qh4 Qxd5 and black has solved the problems - the center has disintegrated, the bishop is proudly developed on f5, and the c5 knight has been exchanged for its counterpart on c3.
MVL tried 13. … a4, with the idea to establish the knight in permanent residence on b3, play … Bg4, … Bxf3, and only then … f5 or … e6, but this is too slow. Carlsen is able to bring an additional defender to the d5 pawn, which removes most of the “sting” from the … f5 break.
We got:
… a4 14. Rd1 Bg4 15. Be2 Nb3 16. Be3 Bxf3 17. gxf3 Qd7 18. h4 e5 19. Nb5 Rac8 20. h5 f5? (diagram)
But this is too weakening, now. Carlsen has a choice of excellent alternatives, and a winning position:
hxg6 hxg6 22. exf5 gxf5 23. d6! (opening the a2-g8 diagonal - there’s a king there) cxd6 24. Qd3 Nc4? (diagram)
Somehow, MVL must not have considered Carlsen’s response, but his position is terrible anyway:
Qxc4+! Rxc4 26. Bxc4+ Rf7 27. Nxd6 resigns. (1-0) - two rooks, a knight and a pawn are more than enough material for a queen…
MVL dropped the first game rather quickly. In the second game, Carlsen sacrificed two pawns and a rook for the bishop pair, and then mounted an assault on the dark squares around MVL’s king until he crumbled under the pressure. Down 2-0 after two games, it did look as though Nakamura’s prediction would come true…
But game three was simply bizarre. Already on move 10 - MVL blunders:
… Qd6? (diagram - white to play)
You can treat this as a puzzle, if you wish - but I don’t think that the correct path was a surprise to MVL, rather I believe he made a simple miscalculation - luckily for the Frenchman, so did Carlsen.
Black’s 10. … Qd6 is simply a blunder, because the queen steps into the path of the oncoming white e-pawn. 11. h6! is obviously threatened, and black would like to play 11. … Bf6, but with the queen on d6, e2-e4-e5 threatens to win material. I believe MVL had calculated 11. h6 Bf6 12. e4 a4 13. Qc2 Nxc3 14. bxc3 e5 and determined that black would be ok here. This turned out to be the game continuation - but both players missed something!
White can play 14. e5! instead of 14. bxc3, because the bishop is absolutely trapped on the f6 square. 14. … Qe6 15. Qxc3 Bxf3 16. gxf3 c5 and, despite his best efforts, black is not undermining the white center in time to save his bishop. 17. Bb5+ Kf8 18. Qc4, for example.
But Magnus missed this move, and played 14. bxc3 immediately. After 14. … e5, which is a standard Grunfeld move, he used almost all his time and then played 15. Bxa6. I’m in no position to transcribe Magnus Carlsen’s inner monologue to you, but I do believe he realized the opportunity he’d missed, one move late, and let it ruin his focus. After 15. Bxa6, white still has an appreciable advantage. And Carlsen played perfectly, but he lost on time 14 moves later, despite an advantageous, or at least equal, position. Just like that, MVL clawed one back in rather unusual fashion. At this point it seemed neither player was in top form.
In Game four - Magnus simply imploded (diagram - Black to play)
Here, obviously the black queen is threatened. After 31. … Qe7 32. b4 white will claim to have a little more space on the queenside, and slightly better pawn structure on the kingside. On a good day, he might even call the a4 pawn and f5 pawn “weaknesses.” But defending such positions - or even squeezing a little advantage out of them somewhere down the line - is a hallmark of Carlsen’s chess. Here he made a really strange decision that is completely uncharacteristic of the world champion.
… Qxd1+?? 32. Bxd1 Rxd1+ 33. Kg2 axb3 34. Qxb3 Rd7 35. a4 g5 36. h4 Kg7 (diagram)
Carlsen had this position in mind when he traded his queen for a rook and bishop. The “point,” such as it is, is that black’s construction looks like a “fortress.” Carlsen intends to shuffle his king back and forth Kg7-g8-g7-g8 and argues that white cannot make progress. On the 31st move, he FORCED this continuation in order to simply secure a draw, rather than defending an equal/slightly worse position for many moves. But this was a bad decision, because this is not, actually, an impenetrable fortress.
a5 Kg6 38. Qb4 Kg7 39. h5 Kg8 40. Kf1 Kg7 41. Ke1 (MVL can put his pieces exactly where he wants them, black isn’t doing anything) Kg8 42. f4 g4 43. Ke2 Kg7 44. Ke3 Kg8 45. c5 Kg7 46. Qa4 Rd5 47. a6 bxa6 48. Qxc6 and white infiltrates… Carlsen resigned 15 moves later.
In 2016, during his title defense against Sergey Karjakin, Magnus famously said, “I don’t believe in fortresses.” That quote gets brought about often, when Carlsen or his opponents employ them. Here, the champ placed his faith in a false fortress, and he was summarily punished.
MVL “woke up” from here and went on to win the match, 13-11. But it was clear that Magnus was not in top shape. After the match, Magnus said that he wasn’t in top form, and then MVL said he was lucky that Magnus was in terrible form.
He would need to do more against Nakamura, if he wanted to win:
Match of the (Mid)Week #2:
Nakamura vs. Vachier-Lagrave (Chess.com 2020 Speed Chess Championship Final)
In last week’s edition we discussed Nakamura’s advantage in the bullet chess. In the final of the 2020 SCC, he made the point for us. After 90 minutes of 5 minute play, 60 minutes of 3 minute play, and 3 one minute games, Naka found himself trailing MVL by a point. He would go on to win the next 5 one-minute games in a row, before taking the match 18.5-12.5. But what makes Naka so special in bullet? He calculates like a machine, and he poses problems that are extremely difficult to “solve” in seconds. Natural moves don’t beat Nakamura, deep understanding of strategic nuances help, but these subtleties are difficult to convert in one-minute chess. Let’s look at the following exchange, as an example:
From the first game of Nakamura’s 5-in-a-row one minute streak that effectively secured him the championship, black to play:
This type of position is exactly the type of position that Nakamura excels in - at least in one minute chess. The position is imbalanced, which creates winning chances. It’s not so important that white is probably better here, with more central space, better development, and access to the a-file. what is important is that black’s moves are easy to find and there are concrete threats available to Nakamura. The space gaining move d5 has secured white some extra territory and bothered Nakamura’s c6 knight, but the knight has been re-routed to the useful c5 square via Nc6-b4-a6-c5, and the long diagonal is open for the dark squared bishop. Nakamura tries to create a little more activity for the g7 bishop:
… b4! (of course)
Now, “correct” for white appears to be 17. Na4 Nxa4 18. Bxa4 Bxb2 19. Rb1 Bc3 20. Bh6! Here black grabs the b2 pawn, and the bishop finds a home on c3 where it protects the passed b4 pawn, but it turns out that 4 moves into the variation, black must sacrifice the exchange, which gives white enough extra material to defend the position. In one minute chess, you aren’t calculating a lot of 4 and 5 move variations. And even if Vachier-Lagrave was aware that the f8 rook is trapped at the end of the variation, it’s pretty obvious that you wouldn’t give your opponent a dangerous protected passed pawn with few seconds on your clock, if you could help it. Vachier-Lagrave looks for a more concrete, forcing solution, but Nakamura keeps the waters muddied:
Bxc5? bxc3!
The simple 17. … dxc5 is fine - but again, Nakamura is trying to create a winning threat - in this case, a passed b-pawn. The computer’s evaluation of the position is not super important here, it’s that Nakamura has a clear win condition and the defense requires precision from MVL. Not so easy in bullet:
Bd4 cxb2 19. Bxg7 Kxg7 20. Qd4+ f6 (diagram)
White to play:
MVL underestimates black’s ability to support the pawn. The a1 rook must move, but it’s important that it stay on the a-file, where it can attack the b-pawn from a2. 21. Ra7! is the computer move, but 21. Ra2 directly is fine as well. The c2 bishop and f1 rook prevent promotion for the time being. And because black can bring extra defenders to the b2 pawn, it’s important that white be able to bring the f1 rook to b1, and potentially even try the knight maneuver Nd2-c4, in order to gather up the offending b-pawn. But MVL makes an intuitive decision - the a1 rook has to move, he places it on b1 where it blocks promotion and hopes to play Rxb2 next move. I wouldn’t doubt it if Naka had this position in his mind’s eye. He’s clearly aware that the b2 pawn can become dangerous:
Rb1? c5! 22. dxc6 Nxc6 23. Qc3 Qb6! (how MVL wishes the f1 rook could come to b1 now…) 24. Rfd1 Ne5 25. Rd2 Be6 26. Nd4 Ba2 and several moves later, MVL resigned.
I think this is really a hallmark of Nakamura’s bullet performance. The “correctness” of the play is not so important. But Nakamura’s win condition is on the board, and his plan is straightforward. The defense requires precision. MVL couldn’t find the right moves in mere seconds, and the match was level. On to the next game:
After 10 moves, Nakamura (with the white pieces) has a little space advantage. White is better here, but this is hardly an advantage in bullet chess. The tension is primarily related to white’s central space advantage. If black can find pawn breaks that destroy the center, he will stand better. If white can improve his pieces while maintaining his central pawns, he will stand better. What’s really interesting here is that Nakamura forces nothing. He simply blitzes out a never-ending stream of small improving moves and deftly sidesteps the one little provocation that MVL is able to make. Shortly thereafter, MVL implodes on his own, making a series of unforced errors:
Rad1 (not the computer’s top choice, but look at what Nakamura does. He simply brings the rooks to the central files. And how can they be bad, there?) 11. … Qe7 12. Rfe1 a6 13. e4 (why not?) cxd4 14. Nxd4 Rfd8 15. Bf1 (The bishop begs for improvement, so it gets improved. This is not brilliant chess or unfathomable calculation, it’s just simple, good moves, played fast.) 15. … Nc5 16. g3 Rac8 17. Bg2 (on autopilot) h5!? (diagram)
MVL “threatens” … h4. The pawn sacrifice is basically a standard maneuver in such positions. It doesn’t carry any particular venom or threat that can’t be addressed - but if we think back to the last game, I think Nakamura’s decision here can be easily understood. Why give the opponent an obvious plan and win-condition (the open h-file)? Nakamura plays instantly, and he plays a move that, while not the computer’s first choice, leaves an impression as a practical decision.
h3!? h4 19. g4!? (diagram)
It’s not necessary to allow the h-file to come open, or defend with the queen along the second rank after something like 18. Qd2 h4 19. f3 Nh5. This is possible. The position is equal, or slightly better for white. But Nakamura simply closes down the king-side and asks, “what now?” The longer he waits, the longer the tension builds in the position (these are very long seconds, naturally) and soon - MVL crumbles under the psychological pressure to “do something useful.” He starts forcing changes in the position, but they are bad changes. Naka does nothing but clean up the mess:
… e5? 20. Nf3 Ne6 21. Bc1 Nh7 22. Nd5 Bxd5 23. exd5 Neg5 24. Nd4! (simply, taking advantage of the pin on the e-file to get the knight to c6) Qf6 25. Nc6 Re8 26. Qd3 e4 27. Bxe4 Qc3 28. Kg2 Nxe4 29. Rxe4 Qxd3 30. Rxe8+ Rxe8 31. Rxd3 and white stands so much better here. It’s easy for Nakamura to fashion a powerful attack with the knight on c6 in combination with the dark squared bishop. But then, MVL probably shouldn’t give up the e-file… 31. … Re2 32. a4 Nf6 33. Bf4 Bf8 34. Bg5 Nd7 35. Re3 Rb2?? 36. Bxh4 Nc5? 37. b4 Nxa4 38. Bf6 resigns 1-0.
After two losses in a row, MVL again found himself defending against the passed b-pawn in this variation of the Ruy Lopez. He’s taking an interesting psychological beating here, because he obviously likes the positions that he’s getting, but he keeps losing them. So why should Naka switch variations or plans?
On move 55, MVL just goofs. 55. Kh2?? (diagram)
… Rxh4+! and Nakamura went on to win.
And here, we’re out of email space for more diagrams. But more of the same was in store for MVL over the next two games. Nakamura plays relentlessly simple and accurate chess without using any clock time whatsoever. Nothing special, but nothing weakening. He plays straightforward plans and generally doesn’t wade into waters where his opponent’s moves are easy to find. Of course, we can also say that MVL was making a lot of “unforced errors” in these games. But this type of baseline tennis from Nakamura is very difficult to sustain for 30, 40, or 50 moves, all made in under a minute. The unforced errors always seem to come, and Naka always seems to punish them.
Earlier in his career, Hikaru played a very different, swashbuckling style of chess. But results speak for themselves. Hikaru Nakamura is your Chess.com 2020 Speed Chess Champion. And he was your 2019 Chess.com Speed Chess Champion, and your 2018 Chess.com Speed Chess Champion…
One more thing!
If you like reading these, please don’t forget to like, forward, subscribe, tweet about it, or whatever you do with stuff you like. It’s free! And if you have questions or comments, don’t hesitate to reach out at JensenUVA@gmail.com or DM on twitter @JensenUVA
Until next time, ARGH! SHAKHMATY!!!