This should be the longest blog post ever written on the Internet. There are so many things wrong with the system that I think I could make about 100 bullet points. So when you read this, please understand that I am intentionally leaving out a huge number of issues and ideas to make things better.
So why exactly does the World Chess Championship System Suck?
A: The tiebreak is extraordinarily stupid:
Most wins, really? Whoever thought of that one deserves a raise. I have an alternative suggestion: How about a tiebreak match?
B: The tournament is often not long enough to result in an impressive winner
Sadly a fourteen game round robin is often going to result in a winner with a score of +2 or +3. I admit that +3 is reasonable, but do you really deserve to play for the World Chess Championship because you scored 8.5 points out of 14 in one single tournament? It’s just not very convincing and very easy to achieve due to one good result. This year we are likely to see a winner with +2.
C: You often end the tournament with players who are out of contention, determining who will play for the World Championship:
A lot of the debate on Twitter has been whether Anand will push very hard tomorrow. Now wait a second you might say, Anand isn’t even in the running. Well nope he’s not, but assuming a draw between the two leaders: if Anand wins, Caruana plays Carlsen for the World Championship, and if he loses or draws, Karjakin plays Carlsen for the World Championship. This is exceptionally stupid.
Another possible scenario would be the leader of the tournament playing whoever is in say 7th place going into the final round. It’s unsporting for someone who is no longer in contention for anything important to determine the World Championship Qualifier.
D: The tournament is nearly like every other tournament you see throughout the year
Another Round Robin, a few extra games but that’s it. What do you notice in these Round Robins? Well they are so short, and such a large % of chess games end in draws, that you generally have a variety of different winners, because there is not enough time to determine who is actually the best player. Topalov won in Norway, Aronian won in St. Louis, while Carlsen won in England. This time probably Karjakin will win. It’s just very random.
If you’re going to determine something as important as the World Championship contender, at least make it different than the tournaments we see all the time that we already know carry huge variance and allow more space for the top players to shine through.
I could go on and on, but let’s move to my suggestions on fixing it? I have lots of different ideas, but let’s start with the easiest one. We already have a format in place, it’s called the Grand Chess Tour.
This is the easiest thing in the world. Invite the same ten players to these event every time and turn it into the World Championship Cycle. Personally I think there should be four of events that span over two years, so let’s work on that assumption.
This results in 36 games. At the end of the 36 games, now that the top players have had plenty of time to assert their dominance, the players with the highest scores after 36 games play a match for the World Championship Title. This of course will usually include the actual World Champion. If they somehow weren’t able to qualify in this format, they certainly don’t deserve to be World Champion.
There are still some issues with this format:
- Someone may win by a gigantic margin but still have to play a World Championship Match on even footing.
- You still have the problem with non contenders facing off with contenders
So how do we fix this? Easy!
Issue one: If you tie the World Championship Match, whoever did better in the cycle wins on tiebreak. I think this would be a reasonable solution.
Issue two: How do you qualify for this Grand Chess Tour to begin with? You need to make the Grand Chess Tour events also serve as qualifiers for the next World Championship Cycle.
For example:
If you finish 3rd or 4th in the four tournament series, you automatically qualify for the next cycle.
If you finish 5th, 6th or 7th: In that case you automatically get to take part in a qualifying tournament for the next Cycle. This tournament will have something like 14 players, will be a Round Robin, and the top five finishers will qualify. Obviously this would be a tournament that the entire chess world would be watching. It’s always great when you have an excuse to run such important events.
If you finish in 8th or 9th: Well now you’re in trouble. You have to actually qualify for the qualifying tournament. That will be a big Swiss tournament (maybe 50-60 players) where the top 9 finishers qualify.
If you finish 10th: You don’t get to play the next cycle no matter what.
Benefits of this System:
- The World Championship cycle is constantly happening throughout the chess year.
- Everyone has something tangible to fight for. Finishing 4th gives you a huge reward compared to finishing in 6th place. People will be actively following the fight for 8th place, therefore almost everyone has something to fight for. Also there is a huge penalty for finishing last, so people will be desperately trying to avoid it.
- The champion loses his enormous privilege. I think it’s pretty justified that if you can’t come in the top two after 36 games, you just don’t deserve to be World Champion.
- There is a very significant sample size. It’s basically a quadruple round robin between the ten most deserving players in the world.
- There are no wildcards and no spots given to those with high ratings. Every player must earn their way in through a very specific play based qualification process.
- I didn’t mention it above, but needless to say there will not be an idiotic tiebreak system. Ties will result in some kind of head to head play.
There are many alterations one could make to a system like this to make it easier or more attractive. For example we could find a way to give some extra privilege to the champion, or we could include more or fewer players, but for now I think that the above is a good start, and will still satisfy those who love slow chess.
I want to close this blog with a few very important points, which I have not touching on in the rest of the blog:
If the competitors of this tournament played 7 games at a time control of G/10 every day, for all fourteen rounds, it would result in a 98 game tournament instead of a 14 game tournament. It’s an ironclad guarantee that the likelihood of the strongest chess player winning this event is much higher than in a 14 game slow chess tournament. The faster the time control is and the more games you have, the better the winner will be. I think it’s useful to understand this if you constantly defend the need for an exceptionally slow time control. When there are many fewer games, even if the time control is much longer, you are making it much more likely for a weaker player to come through, resulting in a less interesting World Championship Match. The top blitz players are nearly always the top chess players.
I am not suggesting that it be a G/10 tournament, but as always I suggest some speeding up of the typical top level chess game so that we can comfortably fit more games in these events.
Second key point: I think Kashimdzhanov’s idea of eliminating the draw as a result is a very good one and worthy of serious examination. I’d make some small alterations but basically the idea is you play, and if the game is a draw, you switch sides and keep playing with the same time on your clock. You could add a few minutes the first time this happens, but eventually someone is going to win every single chess game. This would make chess games so much more exciting, which in my opinion is a very good thing. I would definitely prefer to eliminate the draw as a final result of a chess match. Here’s the link.
So while I like speeding up games and eliminating draws, I think that the Grand Tour idea would be more appealing to the chess purists who want to burn me at the stake anytime I suggest any changes. Also I’d like to give thanks to Yasser Seirawan, as this is all very similar to an idea he told me about in the past.
Greg Shahade 
would carlsen be allowed to play in these? you go on about how its ridicules that anand game tomorrow is a determining factor in who will face carlsen. but carlsen would be in the same position as anand with his results not mattering because he can’t win the right to play himself. ALso from a promotion stand point it would be tough to build up a match if carlsen won the qualifying tournament for the right to face him.
I kind of like your idea of playing a shorter time control. a game in 60 with a 30 second increment. its still a long game, so it wouldn’t hurt people like caruana who maybe aren’t as good at blitz but maybe you can play 10 games a week instead of 5 or 6 games a week.
Playing just blitz is a bit silly because i do think it is a bit different than classical chess.
I agree that 14 games is not enough but i think that 20-25 games is enough.
Another thing while we are on the topic. Kramnik should be in this tournament, why not invite 10 or 12 people. the rules for qualification are not great two from the grand prix (Caruana and naka, two by rating giri and topalov anand as last challenger and two from world cup.
LikeLike
“We already have a format in place, it’s called the Grand Chess Tour.”
Or alternative the FIDE Grand Prix, if it had enough money to attract the best players.
LikeLike
What do you think of Kaufman chess to eliminate draws? Black gets draw odds, but doesn’t have castling rights: main negative(?), throws out most chess opening theory. It’s probably as close to 50-50 as current Armageddons, and you can give time odds (2 hours versus 90 minutes) if it ends up too imbalanced. Of course, my guess is that you’d personally like it to be at fast time controls rather than slow….
LikeLike
I strongly prefer something along the lines of Kasimdzhanov’s idea.
LikeLike
Some critical remarks. I don’t think you really “solve” issue 2, though of course the extra length does (significantly) reduce its probability. It could still happen that someone who is already “fixed” in the standings (particularly with 5th-7th all being equivalent) can decide in the 36th game who finishes 2nd versus 3rd, say.
As a practical matter (which I know is not your main point), given that the GCT has now twice changed their “next-season qualifier” rules (ask MVL and So), and FIDE has been even less reliable in recent memory, the schema would have to survive a couple of iterations before people took it completely seriously.
Onto the more voluminous quibbles:
I am sceptical as to the (“obvious”) claim that the 14-player RR to qualify for the Shahade-GCT would necessarily be a must-see event. It’s essentially an (smallish) Interzonal to old-schoolers, with ostensibly the top 4 missing (already qualified), and really divides into categories: 2 or 3 expected qualifiers, who would ordinarily fight for 1st but now just drift once enough points are made; the middle guys whose GCT-life depends on happening to be 5th rather than 6th (with playoff matches/schemes often needed); and the also-rans (including those just off-form at the start), who can unfortunately influence everything in late rounds. Not the greatest mix in my opinion, though admittedly the creme de la creme are not unduly influenced. From a marketing standpoint, I don’t think it would attract as significant attention in the current marketplace as perhaps hoped, due to the good number of non-cycle events of similar (or greater) stature. Additionally, the large Swiss play-in to the qualifier (50 or 60 to reach 9) would be beset by similar effects, as seen in Manila 1990 and Biel 1993.
I would have to calculate probabilities, but I’m also not sure that 10th versus 8th-9th is that much of a “huge penalty” (even counting the psychological “elimination from WC” aspect), especially if there is enough professional money outside the cycle. If you really are 8th-9th in the Shahade-GCT, your chance of struggling/surviving through 2 qualifying events (including the 5th-7th finishers at the second stage) is no sure thing in the first place. Even if it works out, if you are a player in his fading years (cue Topalov), the playing of 36 games without a decent chance of coming in 1st/2nd is not that palatable in the first place, without appropriate monetary incentive. At some point, you have to be realistic, and (say) if you are ranked #15 in the world and are 35+ years old, realize that surviving two qualifiers and then playing a grueling GCT circuit to qualify for a Championship match is mostly a dream in any case.
I think my main criticisms survive any simple retooling of your immediate proposal, though your proposal on the whole is no better or worse than any other (indeed probably better than most FIDE ideas), and at the end of the day, money will be the decider as to what is done in practice.
LikeLike
Regarding: you switch sides and keep playing with the same time on your clock.
At least in the given link, Kasimdzhanov (2011) never gives this variant of “same time” on the clock. He just says you play a long game (4 or 5 hours), then a 20-minutes-each game with switched colors, then a 10-minutes-each game switching again, etc. I found the “Resume Clocks” idea listed in a Chessbase post (by Gene Milener) from 2008, under point 7b in an appendix labelled Off-the-board proposals, here he rejects it as it “does nothing to improve the long game, the only game we care about.”
I think Kasimdzhanov is way too optimistic in its potential for interest (and also the gratuitous tennis analogies don’t sit well with me), though I agree that if chess is a “true sport” then at the very least the *attitude* of unfought draws needs to change. Given the typically round-robin aspect of chess, he would do better by comparing to sumo wrestling, than the KOs of tennis.
LikeLike
I think the proper citation is Harry Cohen.
http://en.chessbase.com/post/reader-feedback-the-great-draw-debate-continues
Harry Cohen, Ellicott City, MD USA
I’d like to propose a practical solution to the problem of GM draws. I suggest that games be played with a single sudden death time control (e.g. game in 150 minutes) and an increment (e.g. five seconds). When a draw is reached, for whatever reason (agreement, repetition, or 50 move rule), the clocks are paused, the players switch colors, and a new game is started with each player having whatever time remains on his or her clock. This continues until there is a decisive result.
LikeLike
Pingback: Quality Chess Blog » Happy with the candidates
Pingback: Ok the World Chess Championship System isn’t THAT Stupid | Greg Shahade
I disagree with all of this!
LikeLike
“The faster the time control is and the more games you have, the better the winner will be. ” I liked it until this sentence, which just seems obviously false, in a few ways. Did Naka pay you to write this? hehe
LikeLike