World Cup Seeding System

When you have a number of entrants (128 in this case) competing in a knockout format, what kind of seeding/pairing system should be used?

The most common pairing system involves splitting the list in two, so that No.1 plays 65, 2 plays 66 and so on until 64 plays 128. This is seen as normal, although it’s slightly odd that the players ranked from places 60-68 (give or take) might only be separated by a few rating points, yet will have vastly differing chances of making it through the first round, depending on which side of the halfway line they happen to fall.

The actual system being employed at the World Cup involves the top seed playing the bottom seed in each round, i.e. No.1 vs. 128, 2 vs. 127, all the way up to 64 vs. 65. The main argument in favour of this system is that it gives the highest seeds the best chance of making it to the end – but does it stack the odds too heavily in their favour while making it too difficult for those players nearer the middle of the rankings?

This brings us to this week’s poll question: Is the World Cup pairing system fair?

18 thoughts on “World Cup Seeding System”

  1. The most common knockout pairing system is what the World Cup is doing: 1 plays 128, 2 plays 127, etc. If all matches in the first round are won by the favorite, you then get second-round pairings just like what you would expect for a 64-person tournament: now 1 plays 64, 2 plays 63, etc.

    The 1-vs-65, 2-vs-66, etc. method is used for Swiss System tournaments but not generally (to my knowledge) in knockouts.

    One nice thing about the knockout seeding system is that the difficulty of your opponent rises smoothly as your seed gets better, as opposed to the Swiss system where it goes down and down and then suddenly makes the maximum possible jump up. Of course in a Swiss you get chances to come back from a bad start.

  2. We have a system created to favour TV and their demand for the favourites to make it through to the final stages of tennis tournaments. With the qualification system favouring higher rated players already, you could argue that there is something unfair about it.

  3. The system’s been around since before TV: see http://www.slate.com/articles/sports/explainer/2012/03/history_of_bracketology_when_did_people_start_filling_out_ncaa_tournament_brackets_.html

    In any system where the challenges facing players are not evenly distributed, I think it makes sense to give preference to the better players. This both makes it more likely for the best players to meet at the end (which I think is a good thing) and discourages sandbagging.

  4. Why not use the System used in tennis ? The first 32 are seeded so for example 1 against 2 can meet only in final, and the rest is drawing of lots

  5. It’s fair because there are other ways to get to the Candidates…and low rated players get the chance to play top GM’s…(and I told my friend Leitao to stop with the caro kan but he wont listen…lol…back home)

  6. I voted that it is fair because it is the same way in basketball. Of course if you are number 128 you will have to defeat number 1 to make it to the second round, but afterwards you will have his pairings, hence easier than number 2 (at least on paper).

    Cutting it like a swiss system makes no sense at all. In swiss systems we want the best players to play each other to decide which is best, and as it is a short tournament, we need to do it as soon as possible (hence for shorter swiss tournaments there is the accelerated swiss), but in a KO tournament we want them to play each other only at the end of the tournament.

    My main concern about KO tournaments is that usually better players drop out soon (Gelfand anyone?) which is bad for us, spectators, hence why make it even worse?

  7. I think that a fair system should be fair to everyone. In order to qualify for the World Cup, a 2550 player has to make a much higher performance rating than a 2650 player, because of the way the Swiss system works. But once they get there, they are faced with harder odds from the beginning.

    Some people really dislike that. I personally have no cat in the fight. But the facts are not widely known.

  8. @Jacob Aagaard
    Well, from a competitive point of view it could be argued either way. For starters the fact that Henriquez after eliminating Gelfand gets an easy pairing (Granda) makes clear that is not so clearcut as many may present the issue, once you have eliminated an stronger opponent, you get his pairings!

    Obviously the pairings could be randomized each round and it would be equally fair for everyone (except that a 2550 will still have an expectancy of meeting harder opponents than Topalov, as Topalov cannot play against himself after all but can play the 2550 player).

    But how the hell are we going to make a ‘fair’ tournament out of a KO system? KO is unfair by nature, moreover when you have only 2 games to play. For me fair means the best player has the greatest chance to win. A perfectly fair tournament will have always the best player winning, hence the actual pairing, biasing the results in favour of the stronger players is trying to ‘repair’ the lack of fairness inherent to KO systems with short games.

  9. @Gollum
    What does fair mean? To me it means that if you perform better than someone rated 100 points higher over the cycle (qualification and World Cup) you don’t walk away with less money on average.

    Most players do not care who wins the event. This is a professional sport and people care a lot about their pay cheque. To shift money to higher rated players that have performed less well can quite easily be seen as unfair.

  10. Doesn’t the current system ensure that the strong and weak players are distributed as evenly as possible across the table? So that wherever you start, you expect to face more or less the same strength of opposition to eventually win. If you would do it randomly, you run the risk that one half of the table is significantly stronger than the other half.

  11. @Raul
    No one said randomly. What I am saying is that if you perform better as a 2550 player than a 2650 player to qualify, you get a 2725 opponent and they get a 2650 opponent. A Swiss system of pairing or a 32 seeded player system would eliminate this unfairness.

    You can of course also deal with the qualification system, but there it would be much much harder to find a solution.

  12. In a Swiss system a few elo points make the difference of playing against a bottom seeded or a top seeded player. Is that fair?

    The current system is in fact most fair for the players in the middle as they get an opponent of around average strength. It’s unfair for the top seeds (too easy) and the bottom seeds (too hard).

    Additionally, using a Swiss seeding system for KO makes the top half of the table slightly stronger than the bottom half.

    Again, I think the current system distributes the strength of players as evenly as possible across the table, meaning (I think) that the paths to win from anywhere in the table are as similar as possible in terms of the expected strength of the opposition.

  13. I dont Know which world champion (probably Steinitz) said beforehand that he was the favorite to win a (closed) tournament. “Why ?” he was asked . “I have the easiest tournament” he answered, “I don’t have to play Steinitz”.

  14. Jacob Aagaard :
    @Gollum
    What does fair mean? To me it means that if you perform better than someone rated 100 points higher over the cycle (qualification and World Cup) you don’t walk away with less money on average.
    Most players do not care who wins the event. This is a professional sport and people care a lot about their pay cheque. To shift money to higher rated players that have performed less well can quite easily be seen as unfair.

    So your proposed definition of fair is that the better the performance (elo-wise) of the player, the better reward, hence it is not fair that better players are to play against weak players, hence making it easier to go to the next round in comparison with weak (or medium) players.

    I can agree up to a certain point. I mean, one problem we have with chess is that to live a good life as a professional you need to be at the very top, otherwise you don’t get invited where the big bucks are. To get to the very top, you need to play opens where is relatively easy to lose elo (not only because your opponents will be weaker and a slip costs you dearly, but also because you may want to draw with a weak opponent to guarantee a good prize, rather than fight, lose and go home without anything), so at the end of the day only very skilled people who managed to gain enough…

  15. … people who managed to gain enough elo (or with the right nationality, say Germany, Holland, Spain or US now) get invited to super tournaments.

    But trying to solve the professional chess scene with a single knock out tournament is useless, the changes should be addressed at closed tournaments, and maybe we as spectators should care more about the fighting of the games than the average elo of the players, hence allowing 2600 players to be playing in big tournaments. I think the Wijk aan Zee touranment has the right system, having a B group where you can win and classify to the main tournament.

    So at the end of the day, for me the right definition of fair is that the best player has the greater chance to win the tournament hence the actual pairings of this tournament are the right ones.

  16. How about playing the “tiebreak” before the classical games? This would add even more fighting chess, two peacefull draws would then be out of question.

Leave a Comment

Your email address will not be published. Required fields are marked *

Scroll to Top