Added on 12/6/2021

GM Joel Benjamin


Fans, be careful what you wish for. After five games we worried about a repeat of 2018, where all the regulation games between Carlsen and Caruana were drawn. Three games later our blood lust has been satisfied, but now the drama of the ultimate outcome has been largely drained from the match.




Carlsen grabbed the lead in the longest game in World Championship history, 136 moves.  There is a lot to unpack here, with quite a number of critical moments.

Magnus switched back to 1.d4, playing an unusual form of the Catalan.  I think after his concrete approach failed in game four, he opted for a “get a position and play” strategy.


With 10.Nbd2 Carlsen offered a pawn.  Those who listened to my commentary on ICC back in the day know I have long been fascinated at how grandmasters do not hesitate to offer pawns these days.  Black can take this pawn, but White will have nice pressure after something like 10...cxb3 11.Nxb3 Bb6 12.Nfd2.  Ian’s reaction was better—he immediately solved all development issues with an exact sequence, 11…b5 12.Nce5 Nb4 13.Qb2 Bb7.  Nepo has handled opening surprises very well, demonstrating strong preparation, opening instincts, or both.

Normally we would anticipate a neutralization job and a return the next day with the White pieces.  But Nepo proceeded more boldly than we would have expected.  First, he spurned a queen trade on move seventeen, then he encouraged a major unbalancing of the situation.


Here 25…b4 would have given Black a very pleasant position, perhaps a slightly sunnier side of a draw.  But Nepo offered two rooks for a queen with 25….Rac8!?

This move is difficult to evaluate.  In recent years there has been a shift in chess philosophy, with the queen now valued higher against pieces than before.  World-class players make decisions concretely, and the resulting position is difficult to play accurately—indeed, we saw several errors.  My thought is you better be right with such a decision against Carlsen because if you are wrong, he can play with the rooks against the queen for a long time.  And 31…Bb2? was a major error that surprisingly went unpunished.

 32.Rxb2 Qxd3 33.Rbc2 was the way to another boring draw, but White can launch a decisive attack with 33.Rcc2! Bxa3 34.Nf4, abandoning the queenside but zeroing in on the king.  The staunchest defense is 34…Qf8 35.Rd7 Kg8 36.Rdc7! Qb4 37.Rc8+ Kg7 38.Nxh5+ Kg6 39.Nf4+ Kg7 40.R2c7 Qe4 41.Rd8 (or 41.Rd7) and White will collapse Black’s position on his first or second rank.

How did Carlsen miss that? Time pressure is a big part of it; given more time, Magnus would have looked in more directions.  He just wasn’t tuned into that type of continuation.  Carlsen has such great faith in his technical abilities that he thought that approach would work here, but he grossly misevaluated the position after 33.Rd1? which only gave Black winning chances as White’s pieces don’t coordinate that well.

Nepo passed on the obvious move 36…Bxb4, which is also quite strong.  The proof is in 37.Rcc1 Ba3! 38.Ra1 Qg4! and Black holds on to his bishop and the a-pawn, with suffering in the offing for White.  For the third time in the match, Nepo failed to take advantage of a superior position.

So Black ended up with roughly the same position except he had to give up his e-pawn, which gave the White knight a world of extra squares.  There was no apparent way for White to round up the a-pawn, so a draw was still the correct result.

52…Qe4? was a shocking error in a position where almost any waiting move would have sufficed for a draw.  After 53.Rxa3! it was game on.  The position after 53…Bxa3 54.Rxa3 gets a slight nod for White from the engines I ran, but it looks completely lost for Black to me, as White systematically attacks pawns with his knight and rook and takes them off.  So Nepo correctly snapped the h4-pawn, but that was little consolation for losing his pride and joy on the a-file.

In my book World Champion Chess for Juniors, I crowned Carlsen the “Master of Everything and Nothing.”  There isn’t anything he doesn’t do very well, but he is especially adept at “nothing.”  No one sweats the details of technical positions like Magnus.  Here he worked extremely hard, seeing his clock tick down to the last minute or two on many occasions.

I found 59.f3 and 60.f4 intriguing.  Maybe it had to be done to give his king immediate shelter and to drive the bishop from its comfortable perch on the center.  But it would be far from routine maneuvering his rooks with his king exposed.  Magnus was up to the task and made progress, though Nepo threw a good roadblock at him.

Apparently 72…Ba7 was a key mistake, but even here White can’t consolidate his gains because his rooks are tangled up.  But  80.Rxf7+! Kxf7 81.Rb7+ Kg6 82.Rxa7 allowed Carlsen to keep pressing.  The engines again get this wrong, giving White an insignificant advantage.  At some point, the White pawns will come forward, and Black will have to hope that the fortress will crack enough to give him sufficient play against White’s king.

But how exactly to go forward would be beyond most mere mortals.  Carlsen pursued a few false trials before finding a remarkable way forward.

Ah, so the rook block attempts to pin the e-pawn on the a7-g1 diagonal.  Ingenious. 

We reached a crucial moment after 115….Qxh4. For the first time, the position is within range of endgame tablebases.

The verdict:  draw.  But tablebases don’t tell you the practical chances in a position.   I think that most commentators would tell you White’s winning chances are quite formidable, even more so when Magnus is in charge.

The decisive mistake occurred on move 130. Nepo went in the wrong direction—130…Qe6--when apparently b1 and c2 would be drawing squares.  I think the idea is the queen needs to harass the king and knight from behind, to prevent them from coming up the board.  Still, it is worth noting that ordinary engines, without tablebase insight, evaluate all these moves as more or less the same value.

It became apparent almost immediately that Black’s position was hopeless.  Whatever twists and turns we saw before, Carlsen was in his element in the last phase of the game.  When you are able to draw almost at will, one point seems like a very large lead.  Though Nepo actually used the words “no big deal,” we all had to wonder if he would be able to keep his head in the game after such a devastating defeat.



GAME 7  1/2-1/2 DRAW

It was time for the proverbial “calm after the storm.”  The players compelled journalists like me to work overtime on game six, so it is only fair that I take a bit of a rest for game seven myself. 

I have noticed that Carlsen seems partial to certain particular maneuvers in the anti-Marshall.  We keep seeing h7-h6, Rf8-e8, Be7-f8, and the knight seems to always come around to g6, or at least start in that direction.

After 13…Bf8, we were in a new position.  I find this a bit surprising as the moves to this point seem pretty generic.  It might look like White can establish a slight pull, somehow.  But again, nothing materialized, and the players steered to a dead-drawn rook ending, wrapping up around move forty.




Carlsen switched back to 1.e4, taking a shot at the Petroff with 3.d4 rather than 3.Nxe5.

Here Carlsen played 7.Nd2, which curiously, is again the fourth most common move in the database.  White most often removes his king from the open center immediately, but there is also 7.Nc3, offering the knight trade in a slightly different way.  Carlsen’s move is perhaps on the safer side because White doesn’t even accept doubled pawns.

There is a scant high-level experience in this position after 9…0-0 10.Qh5 f5, though the analogous position after the trade on c3 is well-explored.  Nepo uncorked the impressive novelty 9…h5!?

That certainly prevents Qd1-h5…but it also allows a check on the e-file.  Apparently, Nepo was comfortable with 10.Re1+ Kf8.  Carlsen, perhaps because he was already up a point in the match, opted for the seemingly toothless 10.Qe1+.

The obvious 10…Qe7 looks like a whole lot of nothing, but Nepo played by analogy—10…Kf8.  The problem is that after White trades the dark-squared bishops.  White can play without the slightest chance of losing, while Black still has some coordination to work out.

It is true that Black is not in any real danger here.  Still, it was said that Nepo missed something of an opportunity in game five, and I’m not sure Black is happier here than Magnus was there on move 20.  But the next move was a shocker.  We’ve seen moves that may rate as blunders because of missed opportunities, but 21…b5? actually hangs something.

Carlsen thought a few minutes, doing a little due diligence before hitting with 22.Qa3+! winning a pawn cleanly.  The camera feed showed Nepo away from the board at this moment, probably contemplating his blunder and trying to get his head right to continue.  In such situations, one rarely is able to make a successful stand.  Perhaps the best try was 23…Bxh3 24.Qxf7+ Qxf7 25.Re8+ Kh7 26. Bxf7 Bf5 (diagram below)


Maybe Black is not lost but I don’t think he would hold this.  Nepo’s resistance was minimal, and Carlsen enjoyed a lot of extra pawns in a queen ending before Nepo gave up.

Making up two points in six games is an extremely tall order.  Even more so when you consider the losses are fresh, and Nepo has not had time to get steady again.  Factor in that Nepo has not achieved much in the opening and has immediately given away any advantages he has had.  The only real question is what the final margin will be.