My game with Magnus Carlsen

The event

Simul rules

Results

Notes on the notes

The game

Event: Google Simul, 1/16/2014
White: Magnus Carlsen (2872)
Black: Freeman Ng (1799)
Result: 1-0

[Event "Google Simul"]
[Site "Mountain View, CA"]
[Date "2014.01.16"]
[Round "?"]
[White "Carlsen, Magnus"]
[Black "Ng, Freeman"]
[WhiteElo "2872"]
[BlackUSCF "1799"]
[Result "1-0"]
[ECO "C63"]
[EventDate "2014.01.16"]

{C63: Ruy Lopez: Schliemann/Jaenisch Defence} 1. e4 e5 2. Nf3 Nc6 3. Bb5 f5 {
A rarely seen opening because it's not really that good, and a terrible choice for the likes of me to play
against a grandmaster, since it can lead to very complicated, tactical positions,
but it's what I normally play, and I had decided to die as I lived.
Magnus paused briefly before replying with the quietest continuation.}
4. d3 Nf6 5. O-O {This took me out of my book. Oddly, it also takes
Deep Junior, the chess engine I used to analyze this game, out of its book.
Whether this is really a non-book move or not, it seems, along with 4. d3, very Carlsenesque:
he avoids modern cutting edge opening moves that push for every last bit of advantage,
and instead plays simple developing moves that just get him into the middlegame
where he feels his natural strength will be enough to get the win. He's basically
giving up some opening advantage in order to avoid possible opening tricks by his
opponents so the game can be decided by skill alone. And, of course, this is a natural choice
for any grandmaster playing a simultaneous exhibition against much weaker players.} Bc5 6.
exf5 {Carlsen takes the gambit pawn. Deep Junior has him up by 0.48, only half
a pawn, at this point, even though he should be up by that one whole pawn
plus a little more for white's natural advantage for moving first. One way of looking at the discrepancy is that I've
obtained almost a pawn's worth of play from my sacrifice. Will I be able
to do anything with it?} O-O 7. Be3 Bb6 8. Bxc6 dxc6 {So far, according to
Deep Junior, I haven't played any bad moves, but on the other hand, there
have always been multiple good moves available, so the odds were better that
I would happen to play one of them. It rates either recapture here
roughly the same, for example, so there wasn't a wrong choice here.} 9. Nxe5
Bxf5 {Carlsen never tried to hold on to the extra pawn -- indeed, it would have been
very awkward for him to try (which is the idea behind many gambits) -- so I go ahead and take it back now.} 10. Nc3 Qe8 {I was wanting to swing the
Queen over to the kingside, but even as I played the move, I realized that all
the squares it might swing to were covered at the moment, and I
was suddenly worried about having my Queen sitting exposed on the open file. Deep
Junior says I'm still okay, though.} 11. Bxb6 Qxe5 {This was the first time there was
one clearly best move, and I'm happy that it's the one I chose. I would have
lost about a quarter pawn's worth of ground if I had taken the bishop instead.
I don't actually understand why taking the knight was better, especially since
bishops tend to be thought of as being slightly more valuable than knights. At the time, I
thought this might be a way to get my Queen to the kingside, but that turned out
not to be possible after all, so I was just lucky to make the right choice
but for the wrong reason.} 12. f4 {Fending off my queen and any attacking ideas
I might have had. However,
this was Carlsen's first of two subpar moves. Deep Junior says the
weird d4 would have been about a third of a pawn better. I don't get it at all.
It reduces the bishop's options for retreat and loses control of e4. What's
good about it? Let's see how the engine foresees things proceeding...} (12. d4
Qf4 13. Ne2) {Ah, I see now! It was purely tactical. White chases the queen off
the c7-g3 diagonal and wins a second pawn with Bxc7. Well, I can see how Carlsen might
have preferred maintaining containment of my pieces to winning a doubled pawn.
Yet, Deep Junior now has white's advantage at less than a quarter of a pawn.}
12... Qd6 13. Be3 Nd5 {My first subpar move. Deep Junior recommends the
natural } (13... Rae8) {, or } (13... Ng4) { angling to eliminate the bishop, and I did in
fact consider the rook move initially, but then I got it into my head that it
might be nice to force an exchange on d5 to undouble my pawns. I should have
thought more about the landscape that would obtain in the aftermath of that!}
14. Qd2 {Deep Junior says Carlsen should have taken the knight to prevent it from
taking his bishop...} Rae8 {...
and that I should now (of course) have taken the bishop. Instead, I forced him to do
what he should have done in the first place!} 15. Nxd5 cxd5 {Deep Junior says
recapturing with the queen was slightly better - so eliminating the doubled pawns
turned out not to be that valuable a goal after all.}
16. Bd4 {This was the most memorable move of the game for me. At first, I
didn't understand it, and then I realized where the bishop was going! That was
a truly wonderful moment, partly because of the beauty of the smallness of the move
compared to the big effect it will eventually have, and partly because it helped me
to experience a way of looking at chess positions that was a bit more advanced
than I was used to. I think I might have become a
slightly better chess player in that moment!} c5 {
I knew I couldn't stop the next move, and decided my best chance from here on
out was to try drumming up counterplay on the queenside.}
17. Be5 {The big move I saw coming with dread and delight. There's about a third of a pawn
difference in value between this move and, say, just retreating the bishop, and I'm surprised 
it's not much bigger. At the time, I felt like this was a crushing move and the game (or at
least the fun part of the game) was now over for me. The bishop dominates the board, and
*can't ever be dislodged* because I have the wrong color bishop and no knights.}
 Qc6 {I seriously considered sacrificing the exchange here} (17... Re5 18. fe5 Qe5) {just to get rid of the bishop,
but I guess I overreacted, since Deep Junior still thinks I'm in the game. The
MC who was doing commentary mentioned the value of opposite colored bishops
for holding draws in the endgame, and maybe that's why the engine evaluation is not so bad,
but during the game, I was much more worried about the advantage that opposite
colored bishops can give to the attacker! Could this be one of those cases where
the human evaluation differs from the computer's because the human can see that the
bishop can stay there forever and affect every corner of the board, while the computer
only sees that the bishop can stay there as far ahead as it can look? Or is the
bishop just not as powerful as I think?} 18. Rae1 b5 {While I carry on with my queenside plans...} 19.
h3 {...Carlsen begins his own kingside action.} d4 {Carlsen appeared at my
board and I played this move, the second of two instances in the game of a single best
move that I managed to find (though this one's much more obvious to me than
the last one: before I can play c4 to try and open some lines on the queenside,
I have to prevent the d3 pawn from advancing in response) and then, rather than reply instantaneously as he'd been doing all game (except for his very brief pause
after 3. ... f5) he spent about ten seconds thinking things over! It was the
biggest thrill of my chess playing life.} 20. b3 {I'm guessing what he spent
all that time thinking about was whether he needed to do something to slow me
down before getting on with his kingside attack, and I'm oddly pleased that
he decided he did need to do something!} a6 {This was the first of three
essentially random moves I made in the game, in positions where I
needed to do a lot of thinking but before I could even get started Magnus
was standing there and I had to move. (There were only a few of us left at that point,
and Magnus was moving as instantaneously as ever in all the games.) I supposed this does do *some* good in
that it makes both itself and its neighbor pawn safer, but really, it was
essentially a "pass". Fortunately, Deep Junior considers it roughly as good as
several other possibilities. But I was not as lucky the next two times!} 21. g4
Bd7 22. f5 {Carlsen's second subpar move. According to Deep Junior, the best move
was} (22. Qg2) {which forces me
to either swap queens or diminish my own queen's scope and let him line up more heavy artillery behind his pawn
storm. However, the move he played is not bad either, and it had the virtue
of totally confusing me: I flat out failed to see that it
opened the way for Qg5, and went into a little panic trying to figure out
what was going on, and then
suddenly, there was Magnus again at my board and I had to play *something*.} Rd8 $4 {The
losing move because it lets him force an exchange of queens and win
a second pawn in the deal. But even aside from that, what was I thinking, removing
my rook from the open file? My fleeting, partial thought at the time was that my pawn on d4 needed
protection before I could push c4. Here's the very clever alternative suggested
by Deep Junior: } (22... Qd5 23. Qh2 Bc6 {What I didn't notice during the game was that
f5 also weakened the bishop I had once thought an immovable object, so now it
can be threatened, allowing my queen and bishop to reconfigure themselves
into much more active positions. White is still better by that
less-than-a-pawn margin that has persisted through most of the game, but at
least I've still got plenty of play. I doubt I could have found this
on my own, though, even if I had had much longer to think about it. The moment
when Deep Junior showed me this maneuver was the second big educational moment that arose
from this game for me.}) 23. Qg5 {Now white is threatening mate on g7} Qh6 {And I can't guard against it with Rf7 because then he'll take the other rook, on d8 where I just moved it!} (23... Rf7 24. Qd8+) 24. Qxh6 gxh6 25. Bd6
Rfe8 26. Bxc5 {I considered resigning after this, but the player next to
me happened to say at that point that he was about to be mated, and so I
decided to wait until after he lost his game to lose mine!} Bc6 27. Rxe8+ {
Here (as well as on the previous and next moves) Deep Junior says f6 would
have been even better. I was surprised Carlsen let me have the e file.} Rxe8
28. Bxd4 Re2 {And I get a rook to the seventh rank against Magnus Carlsen!
(But all it's going to do is chase the white king where it wanted to go
anyway.)} 29. Rf2 Re1+ $2 (29... Rxf2 30. Kxf2 Kf7 {Okay, Deep Junior, I
believe you when you say that this would have been objectively better, but I'm lost
anyway, and there was no
way I was letting go of my rook. That rook represented my continued presence in the game!
}) 30. Kh2 Rh1+ 31. Kg3 Rg1+ 32. Kf4 Rh1 {Carlsen lets me have a pawn so he can
take over the e file and infiltrate with his rook. I accepted the deal because of, you know,
my irrational attachment to my rook. Moving it back to e1 to resume guarding the e file
would only have resulted in it being challenged by Magnus' rook and traded off eventually.} 33. Re2 Rxh3 34. Re6 h5 {The final
"random" move. I had no idea what to do about the rook's attack, which was way more
powerful than I thought it would be, so when it was time to move, I had no candidate moves
in my head except this one that I'd thought about earlier, in a very different position.} 35. Rxc6 {1-0. I thought
Carlsen had been moving "instantaneously" earlier, but when he saw me play this move, before he
was even fully standing in front of the board, his hand shot out
like a striking rattlesnake to snap off my bishop!} 1-0