General Aspects of Tournament Systems and Tiebreakers
This page discusses some aspects of different tournament systems and, where applicable, the tie breakers used in them. This information may give those wishing to organize a tournament the required knowledge to choose an appropriate system and allows them to consider the strengths and weaknesses of using tie breakers.
|Table of contents|
Generally, tournaments have several purposes. These might include:
- Determining a single winner amongst the participants.
- Ordering players by playing strength.
- Giving players an opportunity to compete.
Several tournament systems are available, each of which has different strengths and weaknesses.
The Round Robin tournament is a tournament where everybody plays against everybody, This system gives players ample opportunity to compete, and is very good at ordering players by strength, though it does not necessarily result in a single winner. In most settings however, round robin is unrealistic, because it requires a number of rounds almost equal to the number of players. For large groups of players, time constraints do not usually allow this. It is mostly popular in small groups or when a lot of time is available (such as in a year long competition).
The Knock-Out tournament system is very good at determining a single winner, and requires a relatively small number of rounds to do this (the 2 log of the number of players). Its drawbacks include that it is not suitable for ordering players other than the tournament winner, and it does not give most players lots of opportunity to play.
The Swiss tournament system is still reasonably good at determining a single winner, and is also quite reasonable at ordering the other participants. All participants will also be able to play in all rounds. If the number of rounds is equal to the 2 log of the number of players, it will determine the winner just as well as knock-out would. If it is larger than that, it becomes more like round robin, in that it becomes better at ordering players, but ties for first place may occur.
The McMahon system is a variant of the Swiss system, which awards players some initial points based on their grade/rating prior to the tournament. This has the drawback that many players will never be able to win the tournament, but the advantage is that it needs even less rounds than Swiss to determine a winner, and also that it eliminates games whose outcome is extremely predictable (such as a 4 dan playing a 7 kyu). Other than that, it has the same advantages and drawback as Swiss. All players can play all rounds, with the added advantage that they will generally play against interesting opponents (of roughly the same strength).
The Double Elimination system is a variant of the knock-out, where a player is only knocked out of the tournament after two losses. It will find a single winner with more credibility than the knock-out, because a player is not thrown out of the tournament for a single unlucky loss. Double elimination has more games per player than the knock-out system, but typically fewer than the Swiss type systems. The double elimination does not give an order for players other than the winner, although it often gets the second place correct.
For informal tournaments that span longer times and that don't require much organisation Ladder tournaments are an option.
In most tournament systems, it is possible for players to be tied, having the same score at the end of the tournament. In some cases, it is desirable to be able to distuingish between these players. For example to determine the single tournament winner or to award an indivisible price to one of them. In such cases, it is necessary to break the tie, and several method for this are available.
Generally, the preferred method is to break the tied by having the tied players play extra rounds, a tie breaker referred to as play-off. When this is not possible, usually due to time constraints, there are several tie breakers available that try to break the tie based on the available data (the games played in the tournament, see tiebreakers).
The predominating opinion is: No. Whenever it can be avoided, do not break ties when determining the final order of players at the end of a tournament. Have players share the same place, and divide the prize money equally among them.
There are exceptions when something cannot be shared: indivisible prizes like titles, seeding places to a next stage of a tournament, or flight tickets.
In many tournament tournaments the most valuable prizes (often money) are issues for the top places, with other players that have done well receiving smaller prizes or access to a book table for choosing a free book. Such prizes are easily shared.
See also: Why break ties
When a tiebreaker cannot be avoided, the question that remain is which one is the most appropriate. This has to be answered afresh for each basic tournament system, purpose of tiebreaker usage, and aims to be achieved by the tiebreaker.
The page on tie breakers contains a listing of possible tie breakers, with a short description of each tie breaker.
- A possible definition copies the definition of SOS: The (current or final) sum of opponents' scores.
- A different possible definition is: Quality of go player, which is indicative of his probability to win a game against other go players.
- "strength of opposition" has been used with different meanings for years on rec.games.go. Rating, rank, tournament score, current tournament score, or others. One should be careful to know exactly which type one is speaking of.
- Instead of "strength of opposition", other phrases like "achievement" or "tournament strength" have been suggested.
- "strength of opposition" is not to be confused with "long-term, general playing strength".
Some tournament directors (of not so big and important tournaments) use things (like pairing programs with their default settings) without or without much thinking, if only it works somehow (i.e., a pairing is made and players are sorted at all). This is not a proof of quality of SOS. Becoming aware of SOS and its background requires a learning curve. Weaker kyu players might look upon SOS as just some numbers without having a clear idea of what they shall mean. The above average tournament director will be able to explain the SOS definition - but knowing or even evaluating its characteristica and comparison with other possible tiebreakers (or sharing places) is another level.
A TD should not use a tiebreaker just because it is available but reflect reasons. If a tiebreaker is recommended, one should look more closely how and why the recommendation is given.
Being applicable in all cases does not equate being useful in all cases.
One can state an opinion that a tiebreaker is "good" or "bad". Such can be meant as either an absolute or a relative quality. As an absolute quality, this has to be referred to a definition of a norm specifying what certain degrees of quality are supposed to mean. As a relative quality, one might compare every two tiebreakers (or ordered combinations of successively applied tiebreakers) with each other and justify that by reasons.
That there are specific advantages does not make a tiebreaker "good" yet. One has to compare them with the disadvantages. Weighting then forms personal opinions.
More specifically, one might define a "good" tiebreaker (under the assumption that some shall be used), e.g., to be significant (in the mathematical sense), contain only useful information (no noise), restrieve all the available information of its own nature, guess the result of a rematch correctly the more often.
- Should one consider "more likely scores as better in a hypothetically continued tournament if the more likely is given due to information collected (also) before the tournament"? No, because once a tournament has started and the players are set into its system, only the results shown during the tournament should matter. Otherwise we should not call the thing "tournament" but "end of a series of evaluated events". For the latter, one could (if one trusts in its meaning and precision) define a deadline and award prizes to players simply in order of their rating at that very moment.
- Should one consider "more likely scores as better in a hypothetically continued tournament if the more likely is given due to information collected during the tournament"? Yes (if ties need to be broken at all). Several tiebreakers fall into this category and so might be considered, e.g., direct comparison and SOS.