After the rest day, the Legends of Chess event entered its Final 4 phase, where the top 4 players play each other in a best of 3 matches format, each match consisting of 4 games played at a time control of 15 minutes + 5 second increment from move 1. The only addition to the previous format is that on a 2-2 tie the players will first play 2 blitz games with a time control of 5 minutes + 3 second increment from move 1, and only then move to the Armageddon. With Carlsen and Nepo taking the first match quite easily, it remains to be seen if Anish Giri and Peter Svidler come back strong in the next two matches.

 

Carlsen 2.5-0.5 Svidler:

I still don't know how g5 works! | Photo: Grenke Classic 2019

Magnus Carlsen can be said to have had many phases in his career. Phase 1 started when he got his rating and continued till he reached around 2700, when he gambled material like Tal in his games and played super aggressive chess. Phase 2 started around 2008-9, when Magnus slowly diversified his chess style, and became more of an endgame specialist, almost never losing and winning long games using his physicality and accuracy. Phase 3 started soon after the London Match in 2018, when Carlsen realised that his style just won't give him the results he desires, and switched to a truly all-round game based mostly on the initiative, combining the daring of his youth with mature technique of his World Championship years. If there is one thing common in all these 3 phases, it is his subtle understanding of the psychological nature of chess. Time and time again he is known to play moves that are very uncomfortable for the opponent to phase, and this match against Svidler was no different.

 

One can either call his decision on move 15 brash underestimation of the opponent, or a brilliant ploy to gain the psychological initiative:

 

Svidler-Carlsen

As Jan Gustafsson remarked in the commentary room during the match, this is not what they teach you when you are improving as a player! Is this utterly brazen move sound? Nope! But did it work? Yes!

Peter Svidler was stunned by Carlsen's attempt to gain squares for his knights, as evidenced by his reaction to the move 15...g5!?:

It was like a deer staring at the incoming headlights of a speedy truck. Peter blundered in just 2 more moves:

 

Svidler-Carlsen

White to play. There is a way to refute Black's play, but it's definitely not the game move17.Rf5? Qe3+!, after which the knights trample White's position. The solution can be found below.

Game 2 saw Svidler create a mess with his sacrifice - a material one that is!:

 

Carlsen-Svidler

This thematic Sicilian Exchange sacrifice is completely justified, and soon resulted in an unbalanced endgame.

Rather surprisingly Carlsen chose the wrong square for his king when Svidler erred and gave him a chance:

 

Carlsen-Svidler

Here following endgame principles, the king should be centralised - 30.Kc1! leaves White with a big plus, and Black is struggling to prove his compensation for the exchange. Instead after 30.Ka2?! Nb4+! Svidler managed to prove his compensation and the game ended peacefully.

Under the pressure to win back a point with the White pieces, Svidler decisively erred move 21:

 

Svidler-Carlsen

It's clear that White has over-extended his position, but it is still not too late to realise that you don't have anything in this position and begin damage control - 21.Bd3! is the only way to keep equalising chances. Instead after 21.d6?! Nd7! here 22.Bd3? is just a blunder as after Nxb2 23.Qxb2 Rc5! Magnus wins both the d6 and f4 pawns, and Svidler resigned on move 27.

Nepo 2.5-0.5 Giri:

I'm not afraid of your opening ideas Anish! | Photo: FIDE Candidates 2020

In a match that saw the longest game of the tournament till now, Nepo had surprisingly high energy levels left, as he came back from the brink to draw Game 1, beautifully won Game 2, and did a Nepo in Game 3 - he pressed his opponent on the clock until Anish blundered big time!

 

Game 1 saw Nepo forgetting his preparation on move 13, leading to grave difficulties:

 

Giri-Nepo

13...Qb6?! was a mistake by Nepo. Instead, 13...Qc7! leads to an equal game, as Nepo proved in Game 3 of the match.

Giri however failed to take his chance:

 

Giri-Nepo

21.Bxf8! Rxf8 22.c4! is the only way for White to retain a significant advantage, with the point being that the d6 pawn is kept alive, and remains a thorn in Black's position. Instead Anish chose to play 21.Rb1?!, which relinquishes most of his advantage, and only managed a draw despite playing on for the next 110 moves(!).

An invigorated Nepo won Game 2 in some style after correctly sacrificing his queen to force a two result game:

 

Nepo-Giri

This was the chance that Giri had to take - 10...g5! is the best move and positionally justified here as White plans to play f4 to open the center. Instead after 10...Nxb3?! the open a-file only added to Black's problems.

Anish soon landed himself in hot soup:

 

Nepo-Giri

16...Nh5 invites a queen sacrifice that Nepo was only too glad to take - 17.Qxh5! Bg4 18.Qxg6 fxg6 19.Bxh6 and it is clear that only White can play for a win in this position. Nepo played a clean game and managed to win on move 40.

Game 3 saw the same Grunfeld line being repeated from the 1st game, but Nepo correctly chose 13...Qc7! and equalised. Giri invested a lot of time, which meant that at the critical moment he could not find the win after a typical Nepo gamble:

 

Giri-Nepo

24...Nc3+? was an unnecessary move by Nepo, when simple chess with 24...a4! would have kept the pressure on White. Here Anish had the chance to consolidate effectively with 25.Bxc3 bxc3 and now not the game move 26.Nc2? Qf6(next diagram) but the surprising 26.Nb5!, after which White is clearly winning.

Time trouble reared its ugly head, as Giri blundered decisively on move 27:

 

Giri-Nepo

27.Qd4! performs damage control and keeps the game level. Instead, after the strange 27.Rh2? Rad8(?!, Bh6! was cleaner) Black is winning by force, and Nepo made no further mistakes in bringing home the point.

Results after Match 1/3 in the Semifinals