Norway Chess 2025 in 7 Graphs

Now that Norway Chess is over, I want to get a visual recap of the tournament, as I have done for other events in the past.

Scores

The scoring is always a bit difficult to keep track of with Norway Chess. The players got 3 wins for a classical win. If their classical game was drawn, each player received 1 point, and they played an armageddon game, where the winner received an additional 0.5 points.

The crossed sections show the points the players scored in the armageddon games.
The first thing that stood out to me was that White was doing much better in the classical games than Black. Gukesh had an especially tough time with Black, only getting one draw against Wei Yi and losing all his other black games.
One effect of the Norway Chess scoring system is that Carlsen and Caruana have the same points from their classical games, even though Carlsen scored +2 and Caruana +1. In the end, Carlsen got ahead by an additional armageddon win.

I’ll only focus on the classical games in this post, so the data for the remaining plots is only taken from them, not the armageddon games.

Better and worse games

I always feel like I get a better overview of the tournament when I look at how many times the players were better and worse. To do this, I divided the evaluation after each move into 5 categories: much better (> +1 according to Stockfish), slightly better (between 1 and 0.5), equal (between 0.5 and -0.5), slightly worse and much worse.

It’s interesting to see that the three “old” players weren’t worse for many moves, whereas Gukesh and Erigaisi seemed to have very up-and-down games. Similar to his showing in Wijk aan Zee earlier this year, Wei Yi didn’t get many better positions in his games.

Getting into good or bad positions is one thing, but the players still have to convert these positions into wins. So I also like to look at the number of wins the players got compared to their better positions.

Both Carlsen and Caruana were better in 7 of their 10 games, but Caruana managed to convert one more of those games into a win. Gukesh converted 4 out of 5 games where he was better into wins, which is a very good conversion rate.

Similarly, I looked at the number of worse and lost games for each player.

This plot shows a big difference between Carlsen and Caruana. While both of them were worse in 3 games, Carlsen lost only one of them, whereas Caruana lost all three. It’s also quite remarkable that Gukesh was worse in 9 out of the 10 games he played.

Engine analysis

Finally, I want to take a look at what the engine thinks about the quality of play from each player.

Amazingly, Nakamura only made 1 inaccuracy, 1 mistake, and 1 blunder in the whole tournament, and Wei Yi avoided making any blunders.
It’s also interesting to look at the accuracy of each player and their opponents. Note that I’m using my own accuracy score based on grandmaster games, where an accuracy of 50% means “more accurate than 50% of grandmaster games”.

As one could have guessed, Gukesh had the lowest average accuracy. But one can also see that the opponents in his games had a lower accuracy than the opponents of any other player. This indicates that his games were more complicated.
The difference between Carlsen’s accuracy and the accuracy of his opponents is also really interesting. He managed to play the most accurate games himself, whereas his opponents struggled a lot more with the positions they got against Magnus.

Finally, I want to take a look at the volatility of the expected score for each player.

After looking at all previous graphs, this shouldn’t be a big surprise. Gukesh had the most volatile games, while Wei Yi kept things calm during the whole tournament. But it’s good to know that my volatility score gives the picture I expected.

Let me know what you think about this recap and if you want to see something additional for future tournaments.

---

If you enjoyed this post, check out my Substack.