[Event "FIDE World Championship 2021"]
[Site "?"]
[Date "2021.11.26"]
[Round "2"]
[White "Carlsen, Magnus"]
[Black "Nepomniachtchi, Ian"]
[Result "1/2-1/2"]
[ECO "E05"]
[WhiteElo "2855"]
[BlackElo "2782"]
[Annotator "samsh"]
[PlyCount "116"]
[EventDate "2021.??.??"]
{[%evp 0,116,23,23,29,2,10,14,10,-12,6,6,18,19,-1,2,14,-13,5,-36,-3,-42,-50,
-50,-50,-42,-42,-42,-42,-22,0,0,0,-24,0,-57,-57,-57,-57,-67,-67,-75,-73,-79,
-75,-85,-47,-82,-59,-103,-26,-46,-46,-46,-46,-46,-30,-43,-53,-43,-35,-35,-35,
-35,-35,-35,-37,-37,-26,-26,0,0,21,21,13,13,65,88,88,88,111,104,104,104,104,
102,102,102,102,102,103,104,102,101,109,109,109,107,112,112,117,115,115,53,53,
53,53,36,36,34,36,15,53,13,53,53,53,21,53]} 1. d4 {Game two of the FIDE World
Championship 2021 followed a very different script than game one. Rather than
a nitty-gritty endgame, the players quickly reached a wild middlegame that was
next-to-impossible to play well. Every other move was a mistake, but this had
nothing to do with the players playing badly and everything to do with the
position simply being complicated! It's really surprising to me that, so far,
both players have really managed to show strong ideas with White as early as
the first game. Normally, Black is equalizing easily at the start of these
matches. I suspect this one will be more bloody than the last couple.} Nf6 2.
c4 e6 {The first minor surprise. Nepo has mostly relied on the Grunfeld in the
past.} 3. Nf3 d5 4. g3 {Magnus chooses the Catalan, and Nepo goes for the old
mainline.} Be7 5. Bg2 O-O 6. O-O dxc4 7. Qc2 b5 $5 {A very enterprising move,
and a suprising practical decision. Magnus was the one who put it back on the
map in his first game, playing Black against the Catalan, after the 2018 World
Championship match, where he only faced 1.e4. I suspect this is what he must
have prepared for that match and likely would have some very deep analysis.} (
7... a6 {This is much more common, and has been seen in endless top level
games. I even chose it in my most recent classical game myself!} 8. a4 Bd7 9.
Qxc4 Bc6 10. Bg5 Nbd7 11. Rd1 (11. Nc3 h6 12. Bxf6 Nxf6 13. Rfd1 a5 14. Qb3 (
14. e3 Bxf3 15. Bxf3 c6 16. Kg2 Qb6 17. Qe2 Rfd8 18. h4 Rd7 19. h5 Rad8 20. Qc2
c5 21. Nb5 cxd4 22. Rxd4 Qc5 23. Qxc5 Bxc5 24. Rxd7 Rxd7 25. Rd1 Rxd1 26. Bxd1
Ne8 27. Kf3 Nd6 28. Nxd6 {1/2 (28) Grischuk,A (2778)-Nepomniachtchi,I (2792)
Moscow 2021}) 14... Bb4 15. Ne5 Bxg2 16. Kxg2 Qe7 17. Rac1 c6 18. e3 Rfd8 19.
Ne2 c5 20. Qc4 Rac8 21. Qb5 Ne4 22. Qd3 Nf6 23. Qb5 Ne4 24. Qd3 Nf6 25. Qb5 {
1/2 (25) Vitiugov,N (2727)-Wei,Y (2725) Lichess.org INT 2021}) 11... h6 12.
Bxf6 Nxf6 13. Nbd2 a5 14. Qd3 Bb4 15. e4 Qe7 16. Ne5 Bd7 17. Ndc4 Rfd8 18. Qe3
Be8 19. h4 Nd7 20. Nxd7 Bxd7 21. e5 Rab8 22. Be4 Be8 23. Qf3 f5 24. exf6 Qxf6
25. Qe3 Bg6 26. Rd3 Bf5 27. Rc1 c6 28. Rc2 Rd7 29. Ne5 Rd6 30. Nc4 Rd7 31. Ne5
Rd6 32. Kg2 Rbd8 33. f3 Bxe4 34. fxe4 Qxe5 35. dxe5 Rxd3 36. Qa7 Be1 37. Qxb7
Rxg3+ 38. Kh2 Rdd3 39. Qc8+ Kh7 40. Qxe6 Rg6 41. Qf5 Bg3+ 42. Kg2 Bxh4+ 43. Kf1
Re3 44. Rc1 c5 45. Rd1 c4 46. e6 Reg3 47. Rd7 Rg1+ 48. Ke2 R1g5 49. e7 Rxf5 50.
exf5 Rg1 51. f6 Re1+ 52. Kd2 Re6 53. f7 Bxe7 54. Rxe7 Rf6 55. Kc3 Kg6 56. Kxc4
Rxf7 57. Rxf7 Kxf7 58. b4 Ke6 59. b5 Kd6 60. Kd4 h5 61. Ke4 g5 62. Kf5 h4 63.
Kg4 Kc5 64. Kh3 Kd6 65. Kg4 Kc5 66. Kh3 Kd6 {1/2 (66) Caruana,F (2800)
-Shankland,S (2714) Saint Louis 2021}) 8. Ne5 $5 {[%csl Gc6][%CAl Re5c6,Rg2c6]
And now this is almost a novelty. There are no correspondence games, and just
15 OTB games, with none played by anyone too notable. But Magnus was still
playing reasonably quickly and confidently. He was surely still in his
preparation.} (8. a4 $5 b4 9. Nfd2 Nd5 10. Nxc4 c5) 8... c6 $1 {[%CAl Gb8c6,
Ga8b8][%mdl 672] This is a typical response. White does not want to take the
pawn.} 9. a4 $5 (9. Nxc6 $2 Nxc6 10. Bxc6 Rb8 $15 {[%CAl Ga7a5,Gb8b5] White
got the pawn back, but has an unpleasant position thanks to Black's extra
space.}) 9... Nd5 {[%CAl Gf7f6,Rf6e5]} 10. Nc3 $11 (10. axb5 {The machine
prefers this one, and I found some truly remarkable lines while poking around,
but they didn't really work for White at the end of the day. Magnus' choice
seems stronger.}) 10... f6 11. Nf3 Qd7 $6 {[%csl Gb8,Gc8,Gd7][%mdl 16] This is
a safe looking move, and a very understandable choice against an opponent who
is clearly still in preparation. Under other circumstances, Nepo might have
considered 11. ..b4 as a feistier alternative.} (11... b4 {This has to be the
critical move, hitting the knight with a gain of tempo, gaining space, and
getting ready for Ba6 next. The position gets very messy very fast, for
example after} 12. Ne4 Ba6 13. Nc5 $1 Bxc5 14. dxc5 $1 {[%csl Rb4,Rc4,Rd5]
[%CAl Gf1d1,Rd1d8,Ge2e4,Rb4a3,Rb4c3,Rc4b3,Rc4d3,Rc5d6] when White is ready for
some e4 and Rd1, and he has obvious compensation for the missing pawn, but the
game is still very messy. I let my computer run for a long time here, and it
eventually settled on equal. I'm sure Magnus had checked it much further and
found some ways to put a lot of pressure on Black's position. We may end up
seeing these lines play out in subsequent games!}) {[%tqu "En","","","","e2e4",
"Now White is taking a lot of space in the center.",10]} 12. e4 $1 {[%CAl
Re4d5,Rf1d1] Now White is taking a lot of space in the center.} Nb4 13. Qe2 {
[%CAl Re4e5,Rd1d8]} Nd3 $5 {[%csl Gc1,Rc4,Rd3] Very ambitious, and I like it!
Nepo clearly is not playing scared, despite being caught in the opening by the
World Champion. Black gets ready for b4 next. If he can pull this off, White
will be worse.} (13... N8a6 {This was a reasonable alternative, but White has
obvious compensation for the pawn after} 14. Be3) 14. e5 $1 {Clearing the
e4-square for the knight.} {[%tqu "En","","","","c8b7","This feels like the
most human move, although the machines prefer pushing f6-f5.",0,"f6f5","This
was a more solid alternative, but I do not mind Nepo's move.",10]} Bb7 $6 {
[%csl Gb7] This feels like the most human move, although the machines prefer
pushing f6-f5.} (14... f5 $1 {This was a more solid alternative, but I do not
mind Nepo's move.} 15. axb5 Bb7 $1 (15... cxb5 $6 16. d5 $1 {Black is rapidly
losing control. After} exd5 17. Ne1 $1 Nxc1 18. Rxc1 Bb7 {[%tqu "En","","","",
"e5e6","Black's position will soon start to crumble. Once Rd1 comes and d5
falls, things could get very bad very fast.",10]} 19. e6 $36 {[%csl Re6][%CAl
Re6d7,Re6f7,Re2e6] Black's position will soon start to crumble. Once Rd1 comes
and d5 falls, things could get very bad very fast.}) 16. Ne1 cxb5 17. d5 {
[%tqu "En","","","","d3c5","Black has things more or less under control, and
the machine claims he is fine. But this feels a bit inhuman to me, and I am
not surprised that Nepo chose another direction.",10]} Nc5 $1 {[%csl Rc5][%CAl
Gc5b7,Gc5e6,Rc5e4,Rc4b3] Black has things more or less under control, and the
machine claims he is fine. But this feels a bit inhuman to me, and I am not
surprised that Nepo chose another direction.}) 15. exf6 Bxf6 16. Ne4 Na6 {
[%CAl Rc6c5,Ga6c7,Gc7d5] White has obvious compensation for the pawn, but now
Magnus took a very aggressive decision. I really doubt he missed any basic
tactics; I suspect he just overestimated his compensation for the exchange.
Still, the position remained totally unclear to the human eye.} 17. Ne5 $6 {
The machines really hate this move, particularly in connection with how Magnus
played down the road, but to the human eye, the position is still totally
unclear.} (17. Nxf6+ $5 {This should have been preferred. After} {[%tqu "En",
"","","","g7f6","",10,"f8f6","",0]} gxf6 $5 (17... Rxf6 $6 {[%tqu "En","","",
"","f3e5","",10]} 18. Ne5 $1 {[%CAl Re5d7]} Nxe5 19. dxe5 $1 $16 {[%CAl Re5f6,
Re5d6] This leaves Black in a very bad way, with Be3 and a rook coming to d6
on the agenda.}) 18. Bh6 {[%CAl Rh6f8]} Rf7 {[%tqu "En","","","","b2b3","White
has excellent compensation for the missing pawn. It will be hard for Black to
keep the position under control, and the machine is already saying he needs to
find some only moves.",10]} 19. b3 $5 {[%CAl Gf1b1,Rb1b8] White has excellent
compensation for the missing pawn. It will be hard for Black to keep the
position under control, and the machine is already saying he needs to find
some only moves.} Nab4 $1 20. Nd2 $1 {[%CAl Rd2e4,Re4c5] And apparently here,
f6-f5 is the only clean equalizer, which I doubt I would even consider. In
practice, it feels super hard for a human to play the black side of this
position.}) 17... Bxe5 18. dxe5 Nac5 {[%tqu "En","","","","e4d6","",10,"f1d1",
"",0]} 19. Nd6 ({Stockfish 14:} 19. Rd1 $2 Nxe4 20. Bxe4 c5 21. Bxd3 cxd3 22.
Rxd3 Qc6 23. f3 c4 24. Rd6 Qc5+ 25. Qe3 Qc7 26. f4 b4 27. Bd2 Rab8 28. Qxa7 c3
29. bxc3 Qc4 30. Qe3 b3 31. Bc1 Rbd8 32. Kf2 Rxd6 33. exd6 Qc6 34. c4 Qg2+ 35.
Ke1 b2 $19) 19... Nb3 $11 20. Rb1 $5 {Magnus may have overestimated his
compensation here, but this is very easy to do. Take the position after Black
goes rook grabbing...} (20. Be3 {This was a fine alternative, and should be
around equal.}) 20... Nbxc1 21. Rbxc1 Nxc1 22. Rxc1 {I was watching without a
machine running around here, and I was naive enough to believe White might
even be better. The first question you always need to ask yourself when
evaluating any exchange sacrifice is \"how good are the rooks\". For now, they
are quite bad! Certainly, neither black rook is any better than the d6-knight.
In the meantime, some combination of Rd1, Be4, and Qh5 looks incredibly
dangerous, as Black's king will be super lonely without the help of the
b7-bishop or h8 rook. The machine just laughs and claims Black is much better,
but in human terms, I am not sure I believe he has any advantage at all, much
less a big one.} Rab8 $1 {[%CAl Rb5b4,Rc4c3] Black needs to activate his rooks.
Opening the b-file would solve that problem.} 23. Rd1 {[%CAl Rd1d8]} Ba8 24.
Be4 {[%CAl Re4h7,Re2h5,Rd1d4,Rd4h4]} c3 $2 {This one is a little mysterious to
me. I guess Nepo wanted to give himself a passed a-pawn, but it's hard to
believe he is ready to go make a queen anytime soon.} (24... g6 $1 {This
should have been preferred. Black prevents Qh5 and could be getting ready for
Qg7. The machines claim he is clearly better, but I'm not totally convinced.
White can play Qg4 to threaten a sac on g6, or h4 instending h5, or just
bishop back to g2 to go Ne4-f6. Lots of ideas, and it is very hard for black
to keep all the balls in the air. It's very understandable why Nepo did not
play this way.}) (24... bxa4 $2 {At some point, the machines suggested this
move, but it's not enough for an edge.} 25. Bxh7+ $1 Kxh7 26. Qh5+ Kg8 27. Rd4
$1 {Black faces a vicious attack, but should make a draw after} Qe7 $1 28. Rh4
Qxh4 29. Qxh4 Rxb2 {[%CAl Ra4a3,Rc4c3]} 30. Qxc4 $11 {[%CAl Gc4a2] I'd imagine
White will have to give a perpetual sooner rather than later.}) 25. Qc2 {
[%CAl Re4h7,Gc2e4]} (25. bxc3 $5 bxa4 26. Qc2 $11 {[%CAl Gd1b1,Gc2a2,Gc2a4]
This also looked totally playable for White.}) 25... g6 26. bxc3 {Apparently
here, Black could have been better by letting White take on g6 and then just
ignoring the bishop. Such things don't really occur to a human.} bxa4 (26... a5
$5 27. Bxg6 Qg7 {And Black is better? Whaddya know. This is not going to
happen in human practice.}) 27. Qxa4 Rfd8 {At this point, despite Stockfish's
insistence on absolute equality, I think White is having a lot more fun in
practice. Around here, I thought Magnus would have real chances to score the
full point, but Nepo kept his head above water.} 28. Ra1 c5 29. Qc4 Bxe4 30.
Nxe4 Kh8 $1 31. Nd6 Rb6 32. Qxc5 Rdb8 33. Kg2 $1 {An important move. White
makes sure not to let a pair of rooks trade.} a6 $5 (33... Qc6+ $2 {This is
too desperate. After} 34. Qxc6 Rxc6 35. Rxa7 $1 {Black will face a real uphill
battle defending this endgame.} Rxc3 36. Nf7+ Kg8 37. Ng5 $16 {White wins a
second pawn, and this will be a nightmare to defend.}) 34. Kh3 $1 {Avoiding
the queen exchange.} Rc6 35. Qd4 Kg8 36. c4 Qc7 37. Qg4 {This allows Black to
bail out into a pawn-down but easily drawn ending.} (37. Kg2 {Perhaps this
would have been a bit more patient. I like White's position, but breaking
through is another story.}) 37... Rxd6 38. exd6 Qxd6 39. c5 {[%CAl Gd6c5]} Qxc5
40. Qxe6+ Kg7 41. Rxa6 $11 Rf8 {Black is holding easily despite the pawn
deficit. 3 vs 2 is so dead that I'm even somewhat surprised Magnus played as
long as he did once the queens came off the board.} 42. f4 Qf5+ 43. Qxf5 Rxf5
44. Ra7+ Kg8 45. Kg4 Rb5 46. Re7 Ra5 47. Re5 Ra7 48. h4 Kg7 49. h5 Kh6 50. Kh4
Ra1 51. g4 Rh1+ 52. Kg3 gxh5 53. Re6+ Kg7 54. g5 Rg1+ 55. Kf2 Ra1 56. Rh6 Ra4
57. Kf3 Ra3+ 58. Kf2 Ra4 {I think Magnus was the clear moral winner of the day.
He really managed to get a fighting position, and while he misplayed it at
some moments and gave Nepo some chances (at least according to the machine), I
don't think he was ever particularly close to losing. This should give him
more confidence that he can take risks, knowing that he is totally capable of
fighting back in complicated positions where things have gone awry and the
computer doesn't rate his chances too highly. His white preparation looks very
practical and impressive so far. It will be interesting to see if he repeats
the same endgame tomorrow. I did not get the impression that Black was easily
equalizing, but he certainly will have studied the positions more and might
have reached a different conclusion. For the last 2 matches, he kept the same
black repertoire throughout, but I somehow suspect he will be the first to
deviate in game 3. We find out tomorrow!} 1/2-1/2