top of page
  • Writer's pictureBruin Sports Analytics

Candidates Chess 2024: Why the Most Accurate Chess Player Doesn’t Always Win

By: Billy Peir


Source: ​candidates2024.fide.com

During the entire month of April 2024, 8 of the top chess players in the world competed for one of the most prestigious opportunities in the chess world: to challenge the current world champion, Ding Liren, for the title of world champion. The tournament featured a grueling, 14-round double round robin in which every player played every other player twice, once with each color. In the end, the tournament came down to one of the closest finishes possible: 4 players within half a point of each other, with 2 pairs of those players playing each other in the final round. In the end, the Indian prodigy, Dommaraju Gukesh, came out victorious, becoming the youngest person to win the prestigious Candidates Tournament ever. This came as a huge upset, as Gukesh was the 6th highest-rated player entering and the least experienced out of all of the competitors. In this article, I will try to analyze what actually happened that allowed Gukesh to win.


The Competitors


Before the analysis, we should first introduce our competitors. Each player had a unique road to the candidates tournament, and their accomplishments should not go unnoticed. Introducing the players will also give context to what happens during the tournament. Each player’s FIDE (official) chess rating, at the time of the tournament, will be shown as well as their qualification method.


Ian Nepomniachtchi: FIDE Rating 2758 (7th Global)

Qualified Through: World Champion Runner-Up


Ian Nepomniachtchi, better known as just “Nepo” or “Ian”, pronounced “yawn”, was the winner of the 2023 candidates tournament, and lost in a heartbreaking tiebreaker to World Champion Ding Liren after tying 7-7 in the 2023 World Championship. Ian entered the tournament as one of the favorites, given his pedigree of accomplishments and his record of strong finishes in big tournaments.







Fabiano Caruana: FIDE Rating 2803 (2nd Global)

Qualified Through: 2023 Chess World Cup


Fabiano Caruana qualified for the Candidates tournament by placing 3rd at the Chess World Cup. Caruana has made every single tournament since 2014, as a testament to his chess prowess. Entering the tournament as the 2nd best player by FIDE rating, and with the best player (Magnus Carlsen) withdrawing, Caruana was the clear favorite to win the tournament.







Hikaru Nakamura: FIDE Rating 2789 (3rd Global)

Qualified Through: 2023 Grand Swiss Tournament


Hikaru Nakamura qualified for his first Candidates bid since 2014 by placing 2nd at the 2023 Grand Swiss Tournament. As the clear fan favorite, through his Twitch Streams, Nakamura is known for his insane defensive skills, memory, and extremely fast play in faster time formats. Nakamura was also a strong favorite to win the tournament, along with Fabiano and Ian.







Alireza Firouzja: FIDE Rating 2760 (7th Global)

Qualified Through: Highest Rating


Alireza Firouzja made headlines as the youngest player ever to reach 2800 FIDE rating, surpassing Magnus Carlsen by almost 5 months. After a lackluster 2022 Candidates, the 20-year old Firouzja seemed ready for a sophomore breakout, qualifying for the Candidates tournament as the highest rated FIDE player who had not qualified through other means, beating out American Grandmaster Wesley So by 2 rating points.






Rameshbabu Praggnanandhaa: FIDE Rating 2747 (14th Global)

Qualified Through: 2023 World Cup


Rameshbabu Praggnanandhaa made headlines as the 2nd youngest player to reach the title of grandmaster, at the age of 12 years, 10 months, and 13 days. Since then, Prag, as he is more well known, has not stopped since. The Indian prodigy became the third Indian player to beat Magnus Carlsen in any time-sanctioned game and became the youngest player ever to reach the finals of the Chess World Cup, qualifying him for the Candidates Tournament. The 18-year-old was expected to play around even for the tournament.





Dommaraju Gukesh: FIDE Rating 2743 (16th Global)

Qualified Through: 2023 FIDE Circuit


Gukesh, as he is more well known, was the youngest player in this Candidates tournament at 17 years old. The other Indian prodigy in this tournament, Gukesh is the third-youngest to hit Grandmaster, the third-youngest to break 2700 rating, and the youngest ever to break 2750 rating. He qualified through the FIDE Circuit, which sums up the 5 best finishes of sanctioned tournaments based on difficulty and placement. Expectations were not as high for Gukesh running into the tournament due to his age and inexperience, but his potential to place high was also noticed.



Vidit Gujarathi: FIDE Rating 2727 (25th Global)

Qualified Through: 2023 Grand Swiss Tournament


Vidit Gujarathi qualified for the Candidates tournament by winning the 2023 Grand Swiss tournament after a dominating showing. As the 4th Indian Player to ever surpass 2700 elo, Vidit is the oldest of the 3 Indian players who qualified for the Candidates. As the 25th highest-rated player globally, Vidit was not rated as highly in the pre-tournament rankings, although with the potential to win games.







Nijat Abasov: FIDE Rating 2632 (114th Global)

Qualified Through: 2023 World Cup


Nijat Abasov entered the Candidates tournament as by far the lowest ranked player in the tournament, roughly 100 rating below the next lowest player. After upsetting 5th-seeded Anish Giri in the 2023 World Cup, Abasov went on to go on a historic run that also had him defeat the 20th-seeded Gujarathi. He placed 4th in the Cup, and with the withdrawal of winner Magnus Carlsen from the Candidates tournament, Abasov was the replacement.






The Results



At the end of the tournament, these were the final standings. Gukesh prevailed with a half-a-point victory over Hikaru, Ian, and Fabiano. Prag finished even, and Vidit, Alireza, and Nijat all finished below equal. From the projections at the start of the tournament, there were a couple of big surprises: First, Alireza, who was projected to be a dark horse candidate to win, struggled mightily. The other was Gukesh, who far outperformed his projected rank of 6th to win the entire tournament. So, did Gukesh really outplay the higher-rated and more experienced favorites of Hikaru, Ian, and Fabiano? Or was there something else at play? To start, we will look at the overall accuracy of each player during the event.



Accuracy and Move Distribution


The accuracy of a player during a game is determined by how close a player comes to playing all of the best moves in a game. Chess engines like Stockfish can get pretty close to the best move for any position in the game, so players are compared to chess bots and their moves in order to determine how well they played. An accuracy of 100 means that they played all of the best moves, while an accuracy of 0 means that the player played the worst possible move in every situation. The graph below shows the average accuracy among all games in the tournament.



Although the accuracies are extremely similar to each other, a difference of 1 percent might mean the difference between a win and a loss in a match. For example, Nijat had a 93.19 accuracy, compared to Hikaru’s 94.83, barely more than a 1.5 average accuracy per game. However, while Hikaru was +3 (won 3 more games than he lost), Nijat was -7, a 10-game difference. What is interesting is that although Gukesh won the tournament, both Fabiano and Ian had higher accuracies, with Ian having almost one percent higher accuracy per game than Gukesh. If that is the case, then shouldn’t either of them have won over him? Another metric to look at is move distribution. On chess.com, there are certain categories of a move that each player may make. The best is Brilliant, which denotes an amazing sacrifice. The next best is Great, which is typically an only move that works or a really pivotal move in the game. Following is Best, which is the top stockfish move. Then, in order, we have Excellent and Good, which aren’t the top moves, but pretty close. An Inaccuracy is a move that isn’t accurate but doesn’t immediately lose the game, while a Mistake often is if the opponent plays perfectly. Finally, we have a Blunder, which often loses the game immediately, and a Miss, which is a failure to capitalize off of an opponent's mistake. The move distribution chart is shown below:



From the chart, it becomes clear that the top four players play a lot better than the bottom four players. They all play best and excellent moves more consistently, make less inaccuracies and mistakes, and make more greats and brilliancies. However, Gukesh slightly lags behind the other frontrunners. He plays less greats and less bests. It seems that Gukesh just played worse than the other players, but he ended up winning the tournament, not the other 3. What happened, then, that allowed Gukesh to win where he objectively played worse than his competitors?


Choosing your Opponents


In high-level chess, both players, if they choose to, can make a draw pretty easily. The best openings, if both players play perfectly, end in draws. In the highest level of chess, players win by getting their opponents out of a known opening. These are what are called sidelines, and they are worse than the known theory. The advantage is that your opponent gets thrust out of their knowledge and has to play on their own. You trade off a better position for better chances at winning, and losing. If you want to win a tournament as big as the candidates, you have to take risks. The trick is choosing which opponents you take those risks against. Looking at the four players competing for first, Gukesh, Ian, Hikaru, and Fabiano, we can compare their records against the top four players and the bottom four players.



Against the other top players, there was one game out of 12 where there wasn’t a draw: Fabiano lost to Hikaru once. Now comparing their games versus the other 4 players:



Now we see the difference in playstyle. Gukesh and Hikaru both played hyperaggressive versus their favored matchups, landing a decision in 6 of the 8 games played. Meanwhile, while Ian had the highest accuracy out of all of the players, he also didn’t win as many games against the weaker players. Meanwhile, Fabiano was in the middle, without any losses and one more win than Ian, but ultimately his loss against Hikaru cost him the tournament. So, if Hikaru and arguably Fabiano played as equally aggressive against the other players, what allowed Gukesh to have more success, even though he had a worse accuracy?


Opponent Accuracy


A possible way to determine if Gukesh had success with catching their opponent off guard is to analyze the accuracy of his opponents’ moves. Since every player plays each other twice, there should be little variation in opponents’ accuracy besides how well the player makes their opponent think. The graph is shown below:



As you can see, Gukesh’s opponents had the worst accuracy among all of the players in the tournament. What is interesting is that although Ian and Hikaru also had a ton of wins, their opponents had the highest accuracies against them. Looking at the move distribution chart, we can see the same trend:



Gukesh’s opponents consistently played worse against them compared to the other top players. His opponents played fewer Best and Excellent moves, and made the same amount of mistakes, if not slightly more than his counterparts. However, although Gukesh’s opponents played worse, it is not certain that it was Gukesh’s play that caused them to play worse. It could be explained by variation in focus during the tournament, and Gukesh just got lucky. Another metric will have to be used in order to supplement these findings.


Getting Your Opponent to Think


When you spend a lot of time on a move, it’s typically because you aren’t sure if a move works or not. Typically, it’s when you are put under pressure by your opponent and you have to be extremely precise, or you lose the game. More time spent on a move correlates very highly to making a bad move, as shown in the graph below on all of the moves in the Candidates Tournament.



The better the move is, generally top players take less time to find it. When players make other players take time to think, especially in the first 20 moves of a game, it means that their opponent has put them in a precarious situation, and they are more prone to making mistakes. On the opposite side, players who play moves faster are typically more accurate, as they are more confident in their moves. The graph for average time spent per move on the first 20 moves is shown below:



Ian spent by far the least time on his moves than any other player. The problem was that he didn’t do a great job of putting enough pressure on his opponents. While his opponents were constantly at time deficits, they weren’t forced to think as much. Meanwhile, Gukesh had the opposite outcome. Although he spent the 2nd longest time on his first 20 moves, more than a minute longer than Ian per move, his opponents spent the most time out of any player in the first 20 moves, leading to them being more prone to making mistakes, allowing Gukesh to capitalize on their errors. 


Conclusion



The Candidates tournament only rewards the winner. When you are pitted against seven of the top players, you have to take some risks in order to win. Although Ian played the most accurately out of all of the players, ultimately he didn’t take enough risk to win the event. Meanwhile, Gukesh, the underdog, succeeded in putting his opponents in pressure situations. He sacrificed accuracy for winning chances, and he converted enough opportunities against his weaker opponents in order to slightly edge out his competition. In high-level chess, you have to find the correct balance between playing the best moves and playing the moves that make your opponents uncomfortable. Players are prepared for the best moves, and it’s only when you surprise your opponent is when you will have winning chances.


While Ian lost by far the least amount of evaluation per move, Gukesh forced the most mistakes out of all of the competitors. In the end, Gukesh found just the right amount of risk to take the title. Will Gukesh have the same odds as beating the world champion that a Hikaru, Fabiano, or Ian might have? Probably not. Gukesh is less experienced and lower-rated, and a prepared world champion will be ready for any trick that Gukesh may have up his sleeve. However, he had the correct strategy to win the Candidates, and that is all that mattered. 


Source: candidates2024.fide.com



Sources:
35 views

Commentaires


bottom of page