Fischer Random Final: So beats Carlsen twice in a day

by Carlos Alberto Colodro
11/2/2019 – Wesley So impressed during day two of the finals at the Fischer Random World Championship as he defeated Magnus Carlsen twice to get a whopping 10½:1½ lead, which means he only needs two more points to take the title. In the match for third place, Ian Nepomniachtchi beat Fabiano Caruana once to get a three-point edge before Saturday's 'fast rapid' and blitz games. | Photo: Lennart Ootes /

ChessBase 15 - Mega package ChessBase 15 - Mega package

Find the right combination! ChessBase 15 program + new Mega Database 2020 with 8 million games and more than 80,000 master analyses. Plus ChessBase Magazine (DVD + magazine) and CB Premium membership for 1 year!


1.a4, 2.a5, 3.a6

The second drawn position of the finals was not as similar to the normal setup as the first one was, except for the fact that the rooks were placed on the a and h-files. The bishops were strangely sitting on the central files, while the king from b1 (or b8) would need to move no fewer than five squares in order to castle short.


Wesley So later noted that 1.e4 is clearly the best move, as Black needs to "fight hard to equalize" after Fischer's favourite opening pawn push. The American started the day with White and saw his opponent going for a French Defence structure. In this setup, however, after 1.e4 e6 2.d4 d5 3.e5 the usual 3...c5, played by Carlsen, is not as sound as in the normal French, as White gets to capture and attack the queen — 4.dxc5 xc5 5.b3 — gaining an edge in development.

Apparently the world champion was in a fighting mood though, eager to recover from his loss in game two. Unfortunately for him, however, So has been playing fantastic chess in this event. The American played accurate positional chess to get ahead after the opening. His play was not perfect, but he was clearly the one calling the shots.

In the last move before the time control, Carlsen, who had been defending tenaciously for quite a while, took a step in the wrong direction:


Black's best alternative here was to capture on g2 with the rook, but apparently Carlsen did not want to go into the rook endgame that would ensue after 40...♜xg2 41.♘xf5 exf5 — the computers give White and edge in this line, but as usual a lot of precision would be needed to convert such ending. Instead, after 40...xh4 White patiently found the way to gobble up the e and f-pawns, using well-timed checks with his rook and knight.

In the end, So had two connected passers against Black's single passed pawn on the other flank of the board — a configuration that would reappear in the next encounter:


Carlsen resigned after 60.b5+.

Magnus Carlsen, Wesley So

The moment Carlsen resigned game three of the final | Photo: Lennart Ootes /

The world champion was trailing by six points. Would he try to cut the losses and play a cool game relying on his superiority in faster time controls? Or would he go all-in to bridge the gap immediately? 

Carlsen chose the latter, and resoundingly so, as he opened his white game with 1.a4, 2.a5, 3.a6, once again allowing his opponent to get the upper hand strategically simply by responding with sound positional moves. At some point, So got chances to create an attack against White's king, but the American decided to trade queens instead, going into a superior endgame. And, for a second time in a row, his queenside passers were the decisive factor:


A b-pawn push was again the final move — 50...b3 and White resigned.

Wesley So

Wesley So seems to be getting along pretty well with the Norwegian fans | Photo: Lennart Ootes /

Thus, So will go into the final day of action with a nine-point lead, needing only two points to secure victory — a win in the first 'fast rapid' game would be enough. Of course, to see Carlsen losing three games in a row is not a common sight, especially in slow time controls: the world champion has not lost a classical game of chess since July 2018! Needless to say, he did not leave the playing hall in a good mood, rejecting to speak with the Norwegian press for a second day in a row, as Tarjei J. Svensen reported.

Commentator Danny Rensch and Wesley So commented:

Rensch: He doesn't like to lose two in a row.

So: He doesn't like to lose at all...even in basketball or soccer.

Fabiano Caruana was in the studio when So got the second victory of the day, and explained:

Everyone has bad days, including Magnus, and one bad day in this event is kind of the end. You usually don't get another chance.

Magnus Carlsen

Time for an epic comeback? — Magnus Carlsen | Photo: Lennart Ootes /

Caruana, in the match for third place, did not take the drastic approach Magnus took after losing the first game of the day, as he did not over-push with White and finished the day with a draw, which means he is trailing Nepomniachtchi by three points before the deciding quick-play rounds. Nepomniachtchi got a fine 48-move victory in his game with White, exchanging down into a winning endgame after having successfully dealt with Caruana's threats during the middlegame. The final position:


The match for third place might be the first one of the event to reach the blitz stage, as long as Caruana gets off on the right foot on Saturday.

Fabiano Caruana, Ian Nepomniachtchi

 Fabiano Caruana and Ian Nepomniachtchi | Photo: Lennart Ootes /

Download the PGN files of the games and examine them in ChessBase: Day 1, Day 2.

Rank Name Score Rating
1 So, Wesley 10½/12 2767
2 Carlsen, Magnus 1½/12 2876
Rank Name Score Rating
1 Nepomniachtchi, Ian 7½/12 2776
2 Caruana, Fabiano 4½/12 2812

Commentary webcast - Finals Day 2

Commentary by Sopiko Guramishvili, Danny Rensch and Yasser Seirawan


Carlos Colodro is a Hispanic Philologist from Bolivia. He works as a freelance translator and writer since 2012. A lot of his work is done in chess-related texts, as the game is one of his biggest interests, along with literature and music.
Discussion and Feedback Join the public discussion or submit your feedback to the editors


Rules for reader comments


Not registered yet? Register

genem genem 11/4/2019 09:48
There are two different aspects to chess960-FRC. (1) Non-classical start setup (2) Constantly changing the start setup. I dislike the constant Changing of the setup. Instead, pick one non-classical setup, and stick with it for a decade or so.

"Discard the 'Random' from Fischer Random Chess!"
fgkdjlkag fgkdjlkag 11/3/2019 09:13
@SevenBySeven, memorization and theory really only began shortly before the era of Bobby Fischer. Opening theory was minimal in the 1800s.

@dumkof, comparing humans to engines is not a good benchmark. Humans will never play at the level of an AI/machine learning program like AlphaZero. This is a fact. If you study artificial intelligence/machine learning you will see that it is the case. The moves are based on implicit and explicit very-long term search horizons, which due to exploding number of variations will never be matched by humans. The "intuition" by the programs is intertwined with this calculation, so one cannot correctly use heuristics to match moves played by Alphazero.
SevenBySeven SevenBySeven 11/3/2019 06:28
Memorization and theory are part of the evolution of the royal game that took at least 1,400 years to evolve. Who are we to adulterate this process? Currently an anti-classical chess revisionism is underway, with FRC leading the incursion. As with any other business chess has more than it's fair share of corruption: cheaters, financial whores, and FIDE. Would it not also be disrespectful to put third base in center field, oh and I don't know, home plate where first base used to be? Indeed, Caissa shall have the last laugh in the Endgame(!) with familiar patterns and loads of theory. And thank goodness for that.
dumkof dumkof 11/2/2019 05:02
Kurt Utzinger, I completely agree with you.
Chess is already complicated enough, so why adding some more chaos. Engines are at least 500 Elo points stronger than the best humans, so there is still a lot of space for us to improve. Humans are far from perfect, even at the classical game. Introducing some new variants is similar to Capablanca's chess on a 10x10 board, which never found recognition.
Daniel Miller Daniel Miller 11/2/2019 04:04
In the first game, Sopiko thought Carlsen was clearly better after 4 moves but mush worse after 7 moves. How can we trust what she says about a position?
fixpont fixpont 11/2/2019 03:41
@Kurt Utzinger: why? it is still chess minus opening theory, i quite like this aspect of random chess that players cant blitz out the first ~20 moves from preparation trying to lure the opponent into a trap
PEB216 PEB216 11/2/2019 03:39
As I mentioned in a previous comment, castling in chess960 is awkward, artificial, and unnecessary. My recommendation: Do away with castling in chess960.
Kurt Utzinger Kurt Utzinger 11/2/2019 02:54
I very much dislike Fischer Random Chess. Therefore, I do not understand the hype around this ugly/inelegant/unattractive Fischer Random World Championship. I do not care who wins there and do replay not a single game. But obviously, many chess players see this differently - have fun!
jsaldea12 jsaldea12 11/2/2019 11:19
This is shocking, completely the reverse happened....I TOLD YOU SO CONGRATS SO!!!!...
Lilloso Lilloso 11/2/2019 10:59
Superb Wesley So !
genem genem 11/2/2019 10:16
This chess960-FRC setup 'RKNBBQNR' has White's two knights start on the same shade as each other, namely both start on dark squares. As I discuss in both of the chess books I have authored, this is an under-appreciated difference. Unlike the one classic setup, this 'RKNBBQNR' has far less inherent "knight opposition". This relief from near total "knight opposition" brings us pleasant variety, and a chance to learn more about what fundamentally how these chess pieces could coordinate with each other in new ways, if we would just let them (by having more chess960 tournaments). In the classic setup, the game's four knights all clash in the four center squares, after the knights all make their first developing moves. But in 'RKNBBQNR' each knight is freer to penetrate across rank 4 and into rank 5, on their second moves: the black knights are less able to inhibit advancing white knights. Thanks. Gene Milener