Rules may need to be altered to match increases in draws and the power of chess engines.
On February 10, 1996, the worlds of chess and supercomputing changed forever. In the first game of a highly publicized six-game series, Deep Blue, a chess-playing computer developed by IBM, defeated Garry Kasparov, the reigning world chess champion.
This marked the first time that a computer program beat a chess world champion in a classical game under tournament regulations. Kasparov would rebound to win the match 4–2, but in a May 1997 rematch, Deep Blue emerged victorious 3.5–2.5.
In perhaps the most famous strategy game of them all, machine had bested man.
Deep Blue’s triumph marked a turning point in artificial intelligence, suggesting that machines were approaching, and even surpassing human thinking. Since then, computers have only continued learning, and are now capable of beating humans in even more complex games such as Go and poker.
Nowadays, it’s safe to say that a human chess player will never reach the level of top chess computers, which are still improving. Players of all skill levels regularly turn to chess engines to study both their own games and databases of recorded games.
The influence of this technology is particularly conspicuous at the top levels of competitive chess. In preparation for tournaments and major events such as the World Chess Championship, players will heavily study the past games and tendencies of their opponents. Often with the aid of “seconds,” which are essentially highly-rated assistants, players will attempt to predict opening lines opponents will use and formulate their own strategies.
During broadcasts of last year’s World Chess Championship between reigning champion Magnus Carlsen and challenger Fabiano Caruana, commentators analyzed the ongoing Championship matches while simultaneously pulling up previously played games featuring similar situations to those currently taking place. All the while, chess engines evaluated the game after each move, determining which player was in a better situation.
The effect of advancements in chess engines on the modern game is evident. Players now memorize as many openings and variations as possible, with all the computing power and training tools at their disposal. And although this has certainly increased the skill of top players, it hasn’t come without a cost. That cost comes in the form of draws.
With higher level study comes higher level play, and as quality rises to new heights, the margins between the best players will become slimmer. And, since most people believe that a game played perfectly by both sides ends in a draw (although chess remains unsolved), this means that over time we should expect to see more draws in top competitions.
This brings us back to the 2018 World Chess Championship. The format used for the match was a best-of-12, where the first player to 6.5 wins. For each game, the players received 100 minutes to act, with an additional 50 minutes after move 40, 15 minutes after move 60, and 30 seconds added on after every individual move. This format is known as classical time control. After at least 30 moves had been made by each side, the players were allowed to mutually agree to a draw at any time.
However, for the first time in World Chess Championship history, all 12 games were drawn, leaving the score at 6–6.
This meant the championship went to tiebreakers, starting with a best-of-four “rapid” games, with 25 minutes for each side and an additional 10 seconds after each move. If the match had remained tied, progressively quicker tiebreakers would have been used, ending with a sudden-death match where the player with white pieces receives more time, but a draw counts as a win for black.
However, that wasn’t necessary. Carlsen won each of the first three rapid games, ending the series and retaining his title. It wasn’t the first time Carlsen was forced to tiebreakers. His previous title defense in 2016 against Sergey Karjakin went to rapid games after both players won a game apiece with the other 10 drawn.
In fact, it seemed Carlsen was more than happy to take the match to tiebreakers in 2018. Despite holding a better position after 30 moves in the 12th game, Carlsen offered a surprising draw, preferring his odds in the tiebreak games.
Carlsen’s decision paid off and was likely the right decision, but it was poorly received and no doubt bad for competitive chess. Intentionally playing for a draw is arguably against the spirit of the game and will only harm fan interest in these elite competitions.
Since the World Chess Championship moved to 12-game, champion vs. challenger format in 2008, 75% of games using classical time control have ended in draws. Over the last two matches in 2016 and 2018, that number rises to above 90%.
As standalone games, many of the draws haven’t been “boring,” but that’s not what matters. When the perception is that every game is a draw, which changes nothing in the course of a series, there is little reason for people to pay attention.
Given that draws are occurring so frequently, that we shouldn’t expect this trend to reverse course, and that playing for tiebreakers has become a viable strategy, it might be time to question the method used to determine a world champion.
Should the format of the World Chess Championship be changed?
Full disclosure — I’m by no means a chess expert — I know how to play, and I play on occasion, but that’s about as far as my chess proficiency goes.
I am, however, someone very interested in the structure of sports and competitions in general, and someone who believes we should always be willing to examine our rules and change them if doing so would lead to a better overall contest.
This can be defined in many different ways, including to make a competition fairer or more entertaining. I believe the World Chess Championship is currently failing to meet its original purposes, and because of this, we should consider changing it.
Private matches arranged between top players for the proclaimed title of world champion have occurred since the mid-19th century. Formal systems for determining the best in the world have existed since 1948.
Over time, the specifics have changed. At some points, there have been one-on-one matches. At other times, there have been tournaments. The number of games played in a match has fluctuated. So, too, has the time control.
But while the details may not have always been consistent, the concept has — the World Chess Championship exists to showcase the highest-level chess on earth and determine a clear world champion. In practice, this means to find the best chess players, have them play enough games to generate a worthy sample size, and give them a reasonable amount of time to play these games so they can perform at or near their best.
These are good objectives. Upon closer examination, though, the World Chess Championship does not achieve them.
For most of chess history, it was assumed that the best player in the world would be a human. Deep Blue’s victory over Garry Kasparov in 1997 forever changed that view. For a few years following that match, it may have been unclear whether man or machine was better, but by now, it has been long established that computers are highly superior.
In terms of determining a clear world champion, the World Chess Championship also struggles. Three of the last five championship matches have gone to tiebreakers, and in 2018, not a single decisive result was reached under classical time control. It is now more common than not to determine the chess world champion through rapid chess, a format with its own world championship entirely.
So, if the World Chess Championship does not crown the best chess-playing entity, and is inefficient in determining even the best human player, why wouldn’t we want to change it?
Let’s evaluate some potential changes. My goal here isn’t to advocate for anything in particular, but to open a discussion of both changes and the factors which should go into the decision-making process.
My first instinct was to look at time control, specifically reducing the time players have to make their moves. The more time players are allotted, the more considered and ultimately better their moves can be. It then follows that forcing them to make quicker decisions will lead to more variance and fewer draws.
So, what would the optimal time control be? Unfortunately, there is no easy solution, as this question forces us to rethink what we want from the World Chess Championship.
If that answer is entertainment, to improve the spectator experience and increase fan viewership, then it would make sense to use extremely tight time control.
In recent years, blitz and bullet chess (even quicker than blitz, often allowing each player only one minute for the entire game) have become very popular online for their intense pace and accommodating nature. Not everyone can set aside an hour to play a match, but anyone can find a few minutes to get their fix.
Some of the most exciting chess I’ve seen has come from watching online bullet tournaments. Even though I don’t understand everything that’s happening, it all moves at such a breathtaking speed that it’s almost impossible to look away.
If the World Chess Championship was played in a bullet format, you could play 100 games within the same time it currently takes to play just one classic game. There would be non-stop action, and such a high sample size of games would almost definitely lead to a clear victor emerging.
That might all be true, but it would still be a terrible idea. The best chess players in the world may be computers, but the World Chess Championship should still strive to determine the best human player. Doing so with a series of one-minute games would be a farce.
Preparation for a match would involve getting used to the time control more than anything else, and it makes studying advanced tactics and endgames, in particular, meaningless because there will never be enough time to put concepts into action. It would teach young and recreational players to exclusively play bullet when no person can become great at chess through lightning-quick games alone.
Even from a viewing perspective, bullet games don’t leave enough time for commentators to teach or explain the players’ thought processes. Most important of all, a winner would, in all likelihood, be established early on in the match, as there is a greater variation in the abilities of players as matches get faster.
Let’s not forget that Magnus Carlsen intentionally took his match with Fabiano Caruana to tiebreakers, even when he held a better position in the 12th game. That was just to reach a four-game match of rapid games. Can you imagine how excited he would have been to play Caruana in 100 bullet games? Surely, the match would lose all interest once Carlsen took a significant lead.
Subverting the game for entertainment purposes makes no sense — after all, chess isn’t supposed to be a spectator sport. However, we can’t ignore these interests, either. Some people believe the solution is to extend the championship match to 16 or even 24 games, as it has been in the past.
The additional games might give players more freedom to experiment, but for the very same reason that each individual game is less important, public interest in the event could wane. At some point, the match becomes too long to hold people’s attention, and it could even be putting too much on the players themselves.
As much as chess wishes money wasn’t a factor, without sponsors and media coverage, major events wouldn’t exist, or at least not on the scale they currently do. Chess owes a lot to its marketable personalities such as Garry Kasparov and Magnus Carlsen.
Playing bullet in the Chess World Championship obviously makes no sense. Rapid and blitz are equally poor options, not least because world championships for these time controls already exist.
Perhaps a better option would be to reduce the current amount of time given to players by a third or a half, shortening games while still allowing plenty of time to think and play high-quality chess. It’s unclear how significantly this would affect the frequency of draws in championship matches or public interest as a whole.
Would people even notice the difference between very long and just long games of chess? I’m not sure, but it’s something that could be looked into.
The concept of chess becoming stale due to a “draw death” as players improve and begin to draw more and more of their games is not a new one. Over time, several top players have proposed modifications to the game to combat draw death. Former world champions José Raúl Capablanca and Bobby Fischer have even created entirely new variants of chess.
Fischer’s version, the more popular of the two, was announced in 1996 as Fischer Random and is also known as Chess960. The game plays exactly like chess, except with the position of players’ home rank pieces (the backline of pieces behind the pawns) randomized. With only a few restrictions on the otherwise randomized order, the “960” refers to the number of possible starting positions.
Fischer Random makes the heavy study of openings found in traditional chess impractical, as there are far too many starting positions to memorize and very limited research has been done on each specific position. This makes Fischer Random more about a player’s raw talent and ability to think creatively in unique situations.
Because of these factors, Fischer Random is one of the more appealing chess variants. In fact, the International Chess Federation (FIDE) recently held the first World Fischer Random Chess Championship.
However, because similarly to blitz and bullet chess, there isn’t much value in studying the game, Fischer Random’s value in terms of competitive play seems limited. It will remain popular, but likely never overtake traditional chess.
Neither will Capablanca Chess, which dates back to the 1920s and features an 8x10 chessboard with two new pieces: the archbishop, with powers of both a bishop and a knight, and the chancellor, which can act as both a rook and knight.
Although Capablanca Chess succeeds in both creating a more complex game and one that with defined pieces and no specific time control can be seriously studied, the fact that this version of chess is played on a different board and with different pieces makes it inaccessible to most, and ultimately a tough sell.
There are many different variants to chess, including Seirawan Chess, Four-player Chess, Losing Chess, and others, but while they are worth experimenting with, the future of the Chess World Championship does not lay with a game merely similar to chess.
Perhaps the way forward isn’t with a head-to-head match in the first place, but rather a tournament. As recently as 2007, the world champion was decided through an eight-person, double round-robin tournament. A Candidates Tournament is still used to determine the challenger for the Chess World Championship.
Tournaments are seen as better for risk-taking, as unlike a head-to-head match, where a player can’t lose unless they lose an individual game, draws alone won’t cut it in a tournament. Once one player has won a game, anyone else looking to compete for the championship has to match it.
Because you don’t have total control over your fate, aggression is encouraged, which, in theory, leads to fewer draws. You can’t play to not lose — you have to play to win.
Perhaps draws could be disincentivized even further by adopting the soccer-style scoring system which gives three points for a win and one for a draw, rather than chess’ current format of one point for a win and one-half point for a draw.
With three points for a win, a player with one victory and two defeats and a player with three draws would each have three points. It would certainly be a game-changer, although there could be a contentious debate over the fairness of this scoring format.
There is also the question of whether or not the reigning champion should be given an advantage in this tournament. Before tiebreakers were used, it was traditionally the case that the reigning world champion would keep his title if the match ended in a draw. Emanuel Lasker once even proposed to Capablanca that he should remain world champion unless Capablanca could beat him by two games.
If the Chess World Championship were to move to a tournament format, would it make sense to give the reigning champion an advantage, perhaps starting with the equivalent of one win? Should they be declared champions in the event of a tie? I don’t think so — a world champion should have to prove that they are indeed the best in the world. However, other people may have differing views.
Finally, there are additional modifications to scoring that could be considered. The first of these would be to, in the case of draws, award a win, or perhaps .75 points to .25, to the player with more time remaining.
I have seen this suggested, although it seems to be flawed for a few reasons. Naturally, this would eliminate draw offers, as players behind on time would never agree to a loss. Instead, it could lead to desperate plays and aesthetically displeasing scenarios where players pointlessly move their pieces about until the 50-Move Rule ends the game.
It also seems unfair, particularly when draws are so common, to establish rules permitting players a certain amount of time to act and then subsequently penalizing them for using that time. This rule would unintentionally lead to games of speed chess with each player’s main objective being to play as fast as possible in a way that avoids a checkmate.
A much more reasonable suggestion would be to award .75 points to a player who puts their opponent into a stalemate, and .25 points for the player who is stalemated. Currently, stalemates count as draws for each side, but some, including Lasker, believe that a player should be penalized for being unable to move their king. Changing the way some draws are scored in competitive matches would open up new strategies and reduce the chances of matches ending in ties.
There are certainly more possibilities to be considered, but these represent many of the more commonly discussed solutions to drawing in chess. It remains up for debate how large of an issue drawing is at the Chess World Championships and other major tournaments. It might not be huge at the moment, but if current trends continue over the next few years, it’ll become a much larger talking point.
The concept of draw death is an interesting one. While computers plug away at improving and solving chess with no end in sight, it’s safe to say we won’t be able to achieve perfect play. However, as we reach the upper limits of human chess potential, the very top talents could reach something of a stalemate.
Connor Groel is a writer who studies sport management at the University of Texas at Austin. He also serves as editor of the Top Level Sports publication on Medium, and the host of the Connor Groel Sports podcast. You can follow Connor on Medium, Facebook, and Twitter, and view his archives at toplevelsports.net.