Top-level Chess:

The Chess World Championship

In two posts, I will outline how I think world championship-level chess events should be formatted. This reflects my own preferences as a fan, my sense of what is fair or proper for determining the best chess player in the world, and my sense of what other chess fans might like. I am not a top-level chess player, and I definitely think that the insight and perspective of actual top-level chess players is critical, as they uniquely understand their tournament schedules, the demands of matches and tournaments, as well as how different formats affect the chances of different players. Nevertheless, having spent a long time in chess, including many years as an event organizer and content creator, I think I have some ideas that could advance such discussions.

One side-note before we begin: I am an advocate of using Fischer Random in nearly all professional events, but that is not something I intend to discuss today. Playing regular chess or Fischer Random would change little about my proposal — only how interesting the events would be for players and fans.

The Classical World Championship Cycle

The crown among crowns of the chess world, the champion determined in classical chess time controls has for years been referred to as “The World Champion” while champions in Rapid, Blitz, Correspondence, etc. are referred to as e.g. “The Correspondence World Champion.” This player will also widely be considered the strongest chess player in the world. Of course, there are many factors in the format, which can dramatically change the chances of different players: from time control, to participants, length of the event, match v. tournament play, etc. It is important to pick a format which is likely to both determine the best players and entertain the most fans.

Typically, the World Championship has been decided in a match, but in my view, this is not necessary. Botvinnik won the World Championship in 1948 in a tournament, and Topalov did so in San Luis in 2006 (and if you believe FIDE, world champions have also been determined in their World Cup knockout events). Typically, qualification for the World Championship has started with tournaments, followed by matches between a handful of candidates. The details have gone through many changes over the years, with different problems and different advantages.

The present system has four major deficiencies:

• Rapid chess is an important component. In the case of a tied championship match, the players play a short rapid chess match to determine the champion. This is far from a theoretical possibility, having come up in the last two matches, meaning that Carlsen became “Classical” World Chess Champion in 2018 without winning a single game in the entire Championship Cycle. The classical world champion must be determined by classical chess.
• Short length of the title match. Matches lasting only 12 games are not long enough to provide a satisfying winner. In such a short match if a player gets ahead by a game, they threaten to quickly draw out the match, thus there is less risk-taking, and overall a somewhat different kind of chess than what top-level players usually play. You also need a larger sample size to know who is the best player in the world.
• A controversial qualification cycle. A short-ish candidates tournament with 8 players and a random tie-breaker rather than a tie-breaking match, does not necessarily produce a clear candidate; the most recent one was also fraught with COVID interruption, turning it into two separate 7-rounders, changing participants at the last minute, pressure on players about whether to participate or not given the circumstances (they shouldn’t have), and a weird system of allocating spots in the Candidates (1 player from one tournament, 2 from another, 2 by rating, 1 by organizer’s choice).
• Very high draw rate. This is boring for most fans and potential fans.

Overall, while many people were of course thrilled by the incredible chess players battling in recent World Championship cycles, there was an unfortunately high number of people who found the 12-draw match between Caruana and Carlsen boring and the Candidates controversial.

My suggestion from a theoretical standpoint is to increase interest by increasing fairness:

• Mix tournament play and match play better in the cycle, so that the World Champion will not simply be the best tournament player or best match player, but someone who has proved themselves in both formats.
• Have a more logical and inclusive qualification cycle, which will be both more fair and more interesting for fans to follow.
• Include the defending World Champion a bit earlier in the cycle to reduce their advantage and give fans more chance to watch them play.
• Increase the length of the final phase of the Championship to make it both a more robust determiner of the best player, and more epic for the fans in suspense.

On to the details! There would be three phases to my bi-yearly World Championship cycle.

1. World Championship Grand Prix

A series of 5 10-round swiss tournaments, open to any player rated over 2600, as well as the top 10 players from the World Junior Championship. Players would be allowed to play in up to four of these events, but not all five. Your score in the Grand Prix would be the sum of your three best scores from these tournaments, plus an additional 1 point for any clear first place. The top 7 finishers in the Grand Prix would earn spots in the Candidates Matches. In case of a tie for the last spot(s), short tiebreak matches could be arranged. These would be ~6 game classical matches, with draw odds to the player(s) with the higher performance rating over the course of the Grand Prix.

To give a sense of how many players in the world would be eligible for the cycle, Praggnanandhaa is currently ranked #208 in the world at 2608. Depending on financial details, you might get 200 players participating in the cycle or far less if players without a legitimate chance of winning choose not to participate. In my opinion, these 10-round tournaments would be thrilling, with a great variety of strong players, including a lot of promising teen talent. If 200 participated, you’d be down to max 12 perfect scores after 4 rounds if you had 0 draws. A 50% draw rate would reduce that to an expectation of about 1 perfect score after 4 rounds. In other words, the number of players would not overwhelm the number of rounds, and you’d have time for some players to lead and others to chase.

Schedule-wise, these tournaments could be held every ~2 months, allowing this phase of the championship cycle to finish in 10–12 months.

2. Candidates Matches

The defending World Champion would join the seven qualifiers for a single round of Candidates Matches. The Champion would be the first seed, and the other players would be arranged in order of their finish in the Grand Prix. (If you wanted to encourage the Champion to participate in the Grand Prix, which I would, they could be offered the higher of the 4th seed or their seeding from the Grand Prix). Pairings would then be 1 v. 8, 2 v. 7, etc. Players would play a single match, with the four winners advancing to the World Championship. The match format I would suggest would be first player to win 3 games. If the matches were chess960, the reduced rate of draws would make first to win 4 or 5 games a slight improvement.

The danger in “first to win x games” matches is that a long series of draws causes the match to extend on endlessly, as in one of the greatest matches of all-time, Kasparov-Karpov 1984. One idea to resolve such a problem is to set a maximum length on the matches of, say, 24 games. At that point, if neither player has won 3 games, the player who had first reached 2 wins would advance to the finals. If neither player has won 2 games by that point, they would both be eliminated, and the player who had lost one of the other matches with the greatest number of wins would advance to the finals in their place, with the further tiebreak that if two players had lost other matches with 2 wins, the player who reached 2 wins in the shortest number of games advances.

Schedule-wise, these matches would last a month, and should have at least 2 months before and 3 months after for players to prepare.

I also think it could be ok to leave these matches open-ended until a player does reach the 3rd win. I believe this is not done because of the Kasparov-Karpov match mentioned above, but let’s consider that what we are trying to avoid was arguably the greatest event in chess history! Additionally, I think practical considerations of the difficulty of securing a fancy expensive venue for an indefinite period pale in the face of the importance of having a fair and interesting format, and the world is full of suitable venues where two people can sit in peace, while their game is observed on the internet.

3. World Championship

For the most important event, I have an exciting hybrid format in mind: a 4-player octuple round robin. Players would face each other with white and black, rotating opponents every two games (accompanied by a rest day). This event has some characteristics of a match — you play eight games with a single opponent — and some characteristics of a tournament — you play three different opponents. To me it is a great final stage that ensures the winner is neither profiting from a good matchup against one opponent, nor surpassing an opponent they can’t beat by beating up on weaker players. The incentive to win or draw games would also be a little less static because it is not simply a 1 v. 1. Combined with the factor of two games to watch each day day instead of one, it would be a much more exciting event to witness for fans.

I would suggest that in case of a tie, the two players could be named co-champions; in general I think this practice is more logical than most of the tiebreaks that are used. If a tie-break must be used, perhaps the two players tied for first could play a match where the first player to win a game is champion.

Schedule-wise, the finals would start half-way into the second year of the cycle, and last 35 days without a tie, and perhaps an extra 10 days in case of a tie.

Note- Alternatives

This describes a world championship cycle, which I personally would love to watch and would look forward to watching. Tastes differ, and in general, I advocate trying out lots of changes and different formats. For one thing, you may discover a new format more compelling than any previous ones. For another, different people get to see their preference, and so over the course of many events, you provide something of interest to more different fans. Finally, some people enjoy variety, so even if their favorite format is X, they may prefer seeing formats X, Y, and Z in three different years.

There is an interesting (and succinct) post on the topic of how to re-format world championship-level chess by Greg Shahade here. He mentions Fischer Random as a plausible option for solving the high draw rate and computer preparation taking the life out of the game, and though he does not choose it as his solution, he makes a strong argument for it. Greg’s format of a nearly classical game followed by rapid and blitz in case of draws is not my personal preference, but it is something I would enjoy watching at least once, and that I feel deserves to be tested.

In keeping with my valuing of variety, my next article suggests a new behemoth event, which I could see taking its place alongside the Olympiad and the World Championship as one of the three most enjoyable for chess fans.