Nakamura defeats So in the Semifinal of the Chess.com Speed Chess Championship ("SCC") 2020
T-S-M! T-S-M! - ok enough of that.
Last week we discussed Wesley So’s amazing abilities, and suggested that he’s one of only a few players in the world who can credibly stake any sort of claim on the title of “world’s best online rapid player,” after his dominant wins in the U.S. Championship, and the inaugural event of the Magnus Chess Tour, the “Skilling Open.” To be fair, this success did not occur overnight. For some years we’ve known that Wesley (and his fellow American rival, Hikaru Nakamura) is well suited to online play at fast time controls.
In both 2018 and 2019, So lost to Nakamura in the final of the chess.com Speed Chess Championship (though it is worth noting, perhaps, that Magnus Carlsen did not participate in that event during either year). But these matches were close, unlike many of Nakamura’s SCC early round matches. He defeated Hou Yifan in 2018 by a cumulative score of 27.5-2.5, and in 2019 he thoroughly dominated Vladimir Dobrov, 27.0-8.0. Haik Martirosyan drew the short straw during the first round of the 2020 event, and lost without a fight: 21.0-5.0.
The Chess.com SCC is a unique event that pits players against each other in marathon matches at incredibly short time controls. The format for the 2020 semifinal included 90 minutes of chess at the pace of 5 minutes/side, +1 second increment, followed by an hour of 3 minutes +1 second increment, followed by a half hour of 1+1 games. In practice, this means that a close match will nearly always be decided in the bullet games, because you can get so many 1+1 games in during 30 minutes, that a disadvantage here is crippling. If, like any sane human being, you’re wondering how on earth 1+1 chess is anything but two players moving the pieces around randomly with no time to think, well - for me, it is. For Nakamura it appears to be the crucible that unlocks his calculation ability at its most super-human level. The man is capable of things like this, for example:
Chat calls that 5head.
In any case, the point is that Nakamura really is one of the best at online bullet play, which, somehow, is a skill that can be honed just like classical chess. Carlsen, Firouzja, and Naroditsky probably round out the top four - although So is no slouch either. And this year, So and Nakamura met in the semifinal of the chess.com SCC, where the winner would earn the right to face the victor of the upcoming Carlsen-MVL tilt on the other side of the bracket.
For So to have a chance at finally getting past Hikaru in an SCC event, he’d have been hoping to latch onto an early lead in the 5+1 and 3+1 time controls, where he’s at least Hikaru’s equal, and then cling to it for dear life during the 1+1 games, while hoping the 30 minutes can end as soon as possible (If it’s any indication of how important the bullet games are to SCC results, Chess.com’s pre-match algorithm suggested that Nakamura was a 73% favorite to win, despite the fact that he’s 34 elo points and 11 ranking slots below Wesley So in classical, over-the-board chess).
Let’s see what he was able to do:
Match of the Week #1:
Wesley So - Hikaru Nakamura, Semifinal, Chess.com Speed Chess Championship 2020
Because these games were played at such time controls, they’re full of blunders. Even the openings chosen by the players are occasionally offbeat, sometimes outright risky lines designed primarily to bewilder the opponent. Sometimes an edge on the clock can be worth a pawn, or more, in blitz and bullet - and the earlier you can get your opponent thinking, the better. But for our purposes, these games aren’t always super educational or instructive to the improving player. This is practical chess and psychological warfare at it’s finest, but it’s not always “perfect” play. So rather than go super deep on one game, I’ve found a number of positions from the match that I found interesting, and illustrative of what the world’s best are capable of, even with no time on the clock.
After draws in the first two games, we got a theoretical duel in the Berlin defense, an opening which both players use almost exclusively against 1. e4. Nakamura played a sideline, with 4. d3 and 5. 0-0, that allows black to equalize rather quickly by exchanging a pair of knights on d4, and the players followed the game: Carlsen - Aronian from the 2015 European Team Championships, which Aronian won with the black pieces in spectacular fashion, sacrificing two queenside pawns for a kingside attack and a rook placed aggressively on the second rank. But on move 12, So deviated from Aronian’s play, opting for a more solid set up, and Nakamura responded in a similarly sedate fashion, by slowly improving his knight and making no committal pawn moves. After 21 moves, we arrived in the following position, Nakamura to move, with the white pieces:
And here, Nakamura begins to go wrong. White has the bishop pair (he brought his remaining knight to f5, and So felt compelled to exchange his light squared bishop for the well placed horse), but black has more central space, and dreams of playing … Rad8, … dxe4, and infiltrating with a rook along the d-file. It’s obvious that Nakamura would like to “complete his development,” which means that the c1 bishop and a1 rook should be brought into the game. The human instinct is to find a square for the c1 bishop, but there isn’t a good one. With time, Nakamura likely would have realized that the a1 rook can be brought into the game even without the dark squared bishop moving, if the b-file is opened. 22. b3! is an elegant solution to white’s problems, because 22. … axb3 23. Bxb3 is good for white - the bishop is much happier on the a2-g8 diagonal, and all of a sudden the black d5 pawn is pinned, putting a stop to his plans. And 22. … Rad8 is discouraged by way of the threat to black’s a4 pawn. Black can try 22. … Bd6 to prevent the dark square bishop from being activated on a3, but after 23. bxa4 bxa4 24. Rb1, white is happy to have improved his position by creating an open file for the a1 rook.
Instead, Nakamura thought for a full minute (they’re only given 5, remember?) and obviously felt pressured to “play a move.” Which he did:
Bg5?! h6 23. Be3?! d4 24. Bd2
This is a sad sequence for the dark squared bishop. By move 24, the computer doesn’t really even distinguish between Bd2 and Bc1, saying either square is equally bad for the dark squared bishop. But it’s pathetic to think that black has been gotten in the moves h6, d4, and been passed the move all so the bishop can round-trip back home to c1. Nakamura puts it on d2, but it’s not doing much there.
… Kh8 25. a3? (25. cxd4 was necessary, here) Ba5! (putting a stop to cxd4, are we sure the bishop is better on d2 than c1?) 26. Red1 Bb6 27. Qf3 Rd6 28. Qe2 dxc3 29. Bxc3 Rad8 (diagram)
Here, the computer says that Nakamura is still equal. But it should be clear that So has a dangerous initiative. At the very least, precision is required for Nakamura here. Without active, accurate play, he will be overrun on the queenside, and the black rooks will do damage along the d-file.
Kh1? This is “not it.”
… Bd4! 31. Rf1 Nd7!
The moves 30. Kh1 and 31. Rf1 are made with the express purpose of playing 32. f4. But in typical fashion, So doesn’t try to stop his opponent’s plan, he allows it, because he’s correctly evaluated that white’s plan is wrong. With 30. … Bd4 and 31. … Nd7, So simply says that after 32. f4 exf4 his pieces will occupy strong outposts on the central dark squares, where they cannot be harassed by white pawns. The white central pawns are firmly blockaded on d3 and e4, and as such, they are permanent weaknesses, which do little more than restricting the movement of white’s own forces. It’s tough to comment on the rest of the game because So’s position is so crushing, he doesn’t have to do anything particularly brilliant. He was playing very quickly, finding normal, good moves, and Nakamura had less than a minute on the clock in a hopeless position. So jumped out to the early lead he was looking for after:
f4?! exf4 33. Rxf4 Ne5 34. Rd1 f6 (patience, everything wins for black, there’s no need to take risks in such positions, and no need to force things to come to a head) 35. Bb1 c5 36. Rc1 Bxc3 37. Rxc3 c4 (diagram)
So exchanges off the dark squared bishops, and cracks open the d-file, in order to allow his rooks to win the game. The not-so-thrilling conclusion was:
dxc4? (losing immediately, but really, what else?) Rd2 39. Qh5 Rd1+ 40. Kh2 Rxb1 41. cxb5 Rxb2 white resigns.
In game four, Wesley looked like he was continuing the strong form that brought him a tournament victory at the Skilling open. After 23 moves in the Sicilian, Najdorf variation, we got the following position:
Here, white is better, at least superficially. White has control over the half-open d-file and the d5 square, while black has been unable to rid himself of the backwards d6 pawn, and the light squared bishops have been exchanged. White’s bishop is unapologetically better than its counterpart, and black doesn’t have much going by way of queenside counter-play.
Tactically, white can simply play 24. Qxe5! which nets a pawn because of the pin on the d-file. If that’s too mundane for So’s taste, white can simply improve his position, because black can’t really do anything. The sacrifice Bxh6!? at the right moment is always hanging over black’s head. But it can be prepared by way of Kb1, Rd3, and Rh1. So just went in for it anyway, which doesn’t look to be quite correct:
Bxh6?! gxh6 25. Qxh6 Bd8! (diagram)
I’ve seen other commentators suggest that “So must’ve missed 25. … Bd8.” I suppose. I really don’t know if he just thought that white had a strong initiative here, or what. Certainly, this could be tricky to defend in a blitz game. But the move … Bd8 doesn’t seem very unnatural to me, rather the contrary. To “defend along the 2nd rank” (or the 7th, if you’re playing with the black pieces) is an extremely typical motif in positions where one side is defending the king’s castled position against a sacrificial onslaught. Here, black’s queen controls all the important squares. Incidentally - this is why Rh1 and Rd3, preparing Bxh6, were to be preferred. If white’s rook was already placed on h1, he’d be delivering mate on h8. By inserting the subtle Rh1, white forces black to defend against the mere threat of the sacrifice, without actually making it. This is one reason that, as they say, “the threat is [usually] stronger than the execution.” Here, objectively, white’s sacrifice is incorrect and he’s thrown away his advantage. But defending such a position can be tricky:
Rd3 Qg7! (here, again, we see why it’s better to play Rd3 and Rh1 BEFORE opening the position) 27. Qh3 Bb6 28. Rg2 d5 29. exd5 e4 30. Rdg3 Re8 31. Kb1 (diagram)
Here, Nakamura is plainly playing for a win. He sees that white cannot defend the f2 pawn against the coordinated onslaught of black’s forces, and he prepares to push the e-pawn to promotion, once the f2 pawn falls.
But 31. Kb1 is incredibly calm and collected play from So, considering the danger. In the following diagram, it’s black to play. He’d like to take the f2 pawn and promote the passed e-pawn. Is that possible? Which way to capture? How to evaluate and how to proceed? There are 32 seconds on Nakamura’s clock:
Correct is 31. … Rxf2! because 32. Qh7+ Qxh7 33. gxh7+ Kh8 (not 33. … Kxh7?? 34. Rh3#) 34. Rg8+ Kxh7 35. R2g7+ Kh6 36. Rg6+ Kh5 37. Rg5+ Kh4 38. Rg4+ Kh3 39. Rxe8 Kxg4 40. Rxe4+ Kf5 leads to an ending where black has a bishop for two pawns, and is probably still playing for a win. I would not be shocked, actually, if Nakamura saw this line and felt it might be difficult to convert, since technically R+B vs. R is a draw if white is able to exchange of black’s remaining pawns. In some lines, white is able to create a perpetual check, but white has to play extremely accurately. Lazy moves like 32. Rxf2? lose immediately: 32. … Bxf2 33. Rg2 e3 34. Qf5 e2! and white’s tricks are not working, since the promotion of the e-pawn comes with check and proves decisive in every line.
Unfortunately, Nakamura went astray:
… Bxf2? 32. Qh7+! (diagram)
Now, white is getting his bishop back after 32. … Qxh7 33. gxh7+ Kxh7 (33. … Kh8 34. Rg8+ and 35. Rxe8) 34. Rh2+ Rh6 35. Rxf2, so Nakamura tried:
… Kf8 33. Rh3 Qg8? 34. Qxb7 Qxg6 and Nakamura resigned without waiting for So to play 35. Rh8+ Qg8 36. Rxg8#
Not a perfect game, but a tense struggle worthy of these two players, and much to the delight of the spectators. Here it seemed So was having his way, jumping out to a 2 game lead through 4 games of the five minute section, where he was pinning most of his winning chances. But it appears that Nakamura was energized by this mistake, and he bounced back quickly. In the fifth game, he had the white pieces and, again, he tried to break down So’s Berlin defense in the Ruy Lopez. This time, Nakamura tried 5. c3 (to prevent 5. … Nd4) and after 19 moves, we got this double edged position:
Here, black has the bishop pair, but the center is not opening any time soon. White has not made a tremendous amount of progress on the kingside, yet, but his pieces are well placed for kingside action. Both knights are over there, the bishop is at home on the dangerous a2-g8 diagonal, pinning the f7 pawn to the black king, and black has made some weakening moves with the pawns in front of his king. For his part, So will be looking to create counter-play on the queenside, where he has the open b-file, the advance a5-a4-a3, and white’s king is castled. Nakamura starts by opening lines on the king-side - hoping to create some issues along the open g-file where the black king and queen have alignment issues:
g3 a4! 21. Bc4 Ng6?
It is sort of understandable that So would like to clarify the situation on the g-file before pressing on with … a3. But still, the move 21. … a3 is best here, simply striking while the iron is hot. I suspect that So would spend a lot of time in such a critical position if this weren’t a blitz game.
Rhg1 Bxg4?
But this is just a positional mistake. It’s clear that So is thinking he can connect his rooks with a gain of tempo and delay the opening of the g-file for another move. He plans to follow up with … a3, and now has ideas like Rfb8. But this decision neglects white’s light squared bishop which is now entirely unopposed. In an ending with nothing else on the board, opposite colored bishops is often drawn because the two bishops cannot challenge the other’s influence over the other color complex. In a tense, middlegame struggle with castling on opposite sides, opposite colored bishops simply ups the ante. The side with the initiative finds it a little easier to gin up a decisive attack, usually, because his opponent has difficulty defending the squares his bishop attacks.
hxg4 a3 24. g5!
Here, So’s sense of danger kicks in. He obviously realizes that his queenside threats are not landing before white is able to open the g-file, and he switches to defense mode. He plans to cut the white bishop off from its influence on the kingside by way of:
… c6!? 25. Kb1 Ra5 (trying to play … d5 and cut off the bishop)
The problem with So’s idea is simply that: it’s not in the spirit of the position. Here, black cannot defend the kingside threats passively. He’s not well positioned to do so, his kingside pawn cover is already ruined, and white is well mobilized on that side of the board. So succeeds in shutting the bishop out of the game (a necessity created by the unnecessary and incorrect exchange, 22. … Bxg4), but the process takes so many moves that his position falls apart. Nakamura doesn’t do anything brilliant here, he just opens the g-file like he’s been planning all along and it becomes clear that black’s position is terrible:
gxf4 d5 27. gxh6 Qxh6 28. exd5 cxd5 29. Bb3 Qxf4 (30. f5 was threatened, but black’s position is impossible to hold, in any case) 30. Nxe5 (what else? It’s there for the taking) Kg7 31. Rg4
And So resigned. It turns out, after all that work to try to stop mate on the kingside, the b4 bishop is just hanging. Nakamura claws one back immediately.
Game six was drawn. In game seven, Nakamura again challenged So’s Berlin defense, and this time, So played a little more solidly on the kingside. After 16 moves, we arrived here, with black (Wesley So) to play:
It can be difficult to consistently equalize with the black pieces. Sure - Wesley has played more solidly on the kingside. but white finds a way to exert a little pressure. Nakamura has exchanged his light squared bishop for the knight on c6, doubling black’s pawns. And he’s succeeded in opening the c-file. White has a slight edge, in my opinion, but more critically (especially in blitz), his position is much easier to play. The plan is obvious: Na5, Be3 (or d2, it doesn’t matter), Rec1, etc. takes a bunch of times on c6, and then black’s troubles still aren’t over because the c7 pawn is weak in the end. This is all rather straightforward. Incidentally, positions like this often arise out of the Catalan opening, another opening with which Nakamura is intimately familiar. In that structure, white plays g2-g3 and Bg2, often with the goal of (depending how black responds) playing Nf3-e5 and exchanging several sets of minor pieces on c6. In the resulting structures, as in this game, passive defense of doubled c-pawns on the half open file is not “pleasant.”
In the notes on the previous game I criticized So for exchanging his light squared bishop for the white knight. And here, I’m prepared to criticize him for not doing so. For one thing, the white LSB is gone, so there’s limited downside. For another, the white plan is obvious, as we’ve already outlined. After 17. Na5 Bd7 it should be clear which piece is bossing the other around. At the present moment, So had the opportunity to simply remove the knight before it reached an attractive outpost, and ease some of the pressure on his position. Passive defense is much more difficult, but this is what he went in for:
… Ra6? 17. Na5 Bd7 18. Bd2 Rad8 19. Rec1
Perhaps, here, So convinced himself that the c6 pawn was never falling without white losing the a3 pawn, and he was ok. The problem is really that black can’t do anything useful, while white can play many improving moves. This game goes on and on without black ever really effecting any change in the structure, and without So being able to eliminate the c6 weakness, or make progress anywhere else. If you’re interested, play the moves out on a board at home - it’s actually kindof entertaining, in a sadistic way. Nakamura doesn’t have to do anything, because So’s forces are all completely passive and restrained:
… Qe6 20. Be3 Be8 21. Nd2 Nd7 22. Ndb3 h5 23. Qe2 f6 24. Rc3 h4 25. h3 Bg6 26. f3 f5 27. exf5 Bxf5 28. Rbc1 Nf6 29. Bg5 Nd5
And Nakamura is “forced” to cash in:
Rxc6 Rxc6 31. Rxc6 Qg6 32. Qd2 (diagram)
Here, white has won the c6 pawn, and the c7 pawn is still a weakness. The h4 pawn is falling at Nakamura’s leisure. But black can still claim that the white knights are a bit offside on the queenside, and if ever the a5 knight moves, the a3 pawn falls. So has a very narrow path to achieve any kindof counter-play that takes advantage of his centralized minor pieces. 32. … Bd3! 33. Nc5 e4! and black is fighting. 34. Bxh4 Rf8, for example, and it’s starting to feel as though black’s pieces are coming alive. If ever Nakamura tries Nxd3, exd3 at least gives So a strong, passed d-pawn. But So had just over a minute left, while Nakamura had nearly two, and he just went down without a fight. It is exhausting to defend such positions:
… Be6 33. Nc5 Qb1+ 34. Qc1 Qxc1 35. Bxc1 Bf7 36. Ne4 Bxb4 37. axb4 Nxb4 38. Rxc7 Rxa5 39. Ng5 Be8 40. Be3 Ra8
The material is equal, but the pawns on e5 and h4 are both weak.
41. Bf2 Nd5 42. Rc5 Nf4 43. Rxe5 Ng6 44. Re4 Bc6 45. Rc4 Bb5 46. Rb4 Bd7 47. Bxh4
And So was able to defend neither of them.
Nakamura did not look back. In the five minute section, the players drew another game, and Nakamura won another one. The 5+1 segment ended with Nakamura ahead 5-4, and he would not trail again, “holding serve” in the 3+1 and 1+1 segments that the name “Nakamura” has become synonymous with.
Well played, Hikaru, and fans of SoloMid will be glad to see their guy in the final. He’ll be facing the winner of Carlsen - MVL. No disrespect to MVL intended, I’m sure, but on his stream after the match with So, Hikaru slipped and just said that he’d be playing Carlsen on Saturday. The games will be at noon ET/9am PT.
Puzzle of the Week #16:
But first, a solution to last week’s puzzle:
1. Nxe6! dxe6 (forced, because black can’t deal with both threats, Nxc7+ and Nxg7+) 2. Qxe6! when 3. Bg6+ cannot be stopped. My opponent tried: 2. … Nd8 3. Bg6+ Kf8, but I had calculated this out: 4. Qxe7+ Kxe7 5. Bxh5. White is up two pawns, but he also has the initiative, with the black king stuck in the center. I did go on to convert for a full point.
This week, I offer you a quaint (and very well known) study, supposedly composed by the one and only, Paul Morphy - pride and sorrow of chess - at the tender age of 10 years old:
White to play and mate in two, since there’s nothing sexy about grinding down a material advantage in the endgame.
If you have a solution - email me: JensenUVA@gmail.com, or DM on twitter @JensenUVA
If you enjoy the newsletter, please do subscribe, it’s free! Or like, share, forward to your friends, and all that stuff - it’s fun!
Until next time, ARGH! SHAKHMATY!!