Today’s results were startling. In the classical games, two 2700s were ousted: Maxime Vachier-Lagrave and Daniil Dubov. Today, seven more hit the exits: Wei Yi (who lost to Vasyl Ivanchuk - “old” people can play chess, too!), Francisco Vallejo Pons (who lost to Andrey Esipenko, who has been a 2700 as well - this was a minor upset), Anish Giri (a huge upset, especially since he defeated Nijat Abasov in the first tiebreak game - more on this below), Yu Yangyi (to former 2700 player Etienne Bacrot), Teimour Radjabov (this was a major upset, losing to Jaime Santos Latasa), Le Quang Liem (losing to former FIDE World Champion Ruslan Ponomariov - a moderate upset), and Nikita Vitiugov (a significant favorite against former World Junior Champion Ferenc Berkes).
Back in 1999, Garry Kasparov referred to several of the underdogs still alive in the later rounds of the FIDE Knockout World Championship as “tourists”. That sort of crack was unfortunately characteristic of Kasparov in those days, and in any case what we’re seeing now is that all the players could be seen as participants in a lottery rather than a contest of pure skill. Obviously skill plays the greatest role, but given the high skill level of all the remaining participants and that it takes just one bad game to be eliminated from the tournament, no one is safe - not even Magnus Carlsen.
The biggest upset of the tiebreaks, as suggested above, was Giri’s loss to Abasov. Giri won the first g/25’+10”, but Abasov put heavy pressure on Giri to win the rematch. Abasov squandered some chances in that game, but Giri made the last mistake, allowing the match to proceed to the 10’+10” games. Both games were drawn speedily, and then it was time for a pair of 5’+3” games. In the first, Abasov got nothing with White and the game was another quick draw. With White in the rematch, Giri obtained a small but pleasant and seemingly “eternal” edge - the sort of advantage that a player can milk for a long time with some chances to win and almost no chance of losing. Emphasis on “almost”. Unexpectedly, Abasov outplayed Giri from that inferior position, obtained the upper hand, and then when Giri played the terrible 26.b4 he got rolled. Abasov had a winning advantage after his 31st move, but proving the win might have been difficult. Fortunately for him, Giri froze at the board and lost on time trying to make his 32nd move.
An even more dramatic match in which the favorite barely survived was the one between Vidit Gujrathi and Matthias Bluebaum. The match and the games themselves often went back and forth, and whichever way it would have gone the loser would have been despondent about his missed chances. In the end, Vidit won the first sudden death 3’+2” game in a topsy-turvy affair. First, he outplayed Bluebaum to reach a winning ending with a time advantage to boot, but his opponent’s frantic defense paid off by a miracle when Vidit played 62.Kh6?, allowing Black to achieve a drawn position. Unfortunately for the German, he returned the favor with 64…Kd2??, losing a key tempo and allowing White to reach an elementary ending with a queen vs. central pawn on the 7th rank.
I haven’t had time to analyze all the games I would like to, so here’s a selection from the first tiebreak games of the day. And here are the pairings for round 4, in bracket order (the rating favorite is listed first):
Carlsen - Keymer
Ivanchuk - Sanal
Wang Hao - Svane
Gukesh - Esipenko
Svidler - Abasov
Salem - Vocaturo
Vidit - Bacrot
Nepomniachtchi - Sarin
Caruana - Robson
Duda - Maghsoodloo
Dominguez - Wojtaszek
So - Sarana
Grandelius - Santos
Erigaisi - Sindarov
Ponomariov - Berkes
Nakamura - Praggnanandhaa
In categorizing upsets, I wouldn't attach as much significance to whether a player has (just a bit) more than 2700 or not, and would consider the past and possible future in addition to current ratings. You do the latter for (implicitly) Ivanchuk, Esipenko (who was a past and may be again a future 2700+), Bacrot and Ponomariov (still doing well in KO events).
So may main quibble would be about considering Vitiugov a significant favorite against Berkes. Berkes had also been (barely) 2700+ back in 2011/2013, and had 2680 as recently as December 2021 - then maybe struggling post-pandemic to drop to 2615. It may also matter that he played much more recently than Vitiugov. Berkes had eliminated good ol' Gelfand before - two consecutive wins against higher-rated players are no longer random or reminiscent of a lottery?
A bigger upset may have been Vahap Sanal eliminating former borderline 2700 player Korobov, having eliminated Abdusattorov. Politics and how each player is affected may also matter for both Ukrainians (Korobov) and Russians (Vitiugov, also Esipenko who went from about 2720 down to about 2680 since April 2022).