Executive Summary

The Swiss-system is the standard format for large chess tournaments. It allows hundreds of players to compete without requiring a full round-robin schedule and has proven to be remarkably practical and scalable.

However, several structural limitations of the Swiss-system are often misunderstood or overlooked. These issues affect both the pairing procedure and the interpretation of the final standings. In particular:

• Pairings and rankings are circularly linked. Each round’s pairings depend on the current ranking, while the ranking itself depends on previous pairings. As a consequence, early pairings can have a lasting influence on the final standings.

• Players do not face comparable opposition. Two players finishing with the same score may have faced opponents of very different strength.

• Final standings rely heavily on tie-break systems. Because many players finish with the same number of points, the winner is often determined by secondary criteria such as Buchholz or Sonneborn–Berger. These criteria are only indirect indicators of performance and may produce outcomes that appear difficult to justify from a sporting perspective.

• Colour allocation cannot always be balanced. Even with complex pairing rules, systematic colour imbalances may occur.

• The scoring system is conventional rather than fundamental. Although the classical 1–½–0 scoring scheme has convenient properties, it remains only one possible way to encode game results and does not uniquely determine how tournament performance should be evaluated.

• Floater mechanisms introduce additional asymmetries. The need to move players between score groups (“floaters”) in order to complete pairings can create structural imbalances and add significant complexity to the pairing algorithm.

These limitations do not invalidate the Swiss-system, which remains an efficient and practical tournament format. However, they suggest that some widely accepted assumptions about Swiss tournaments deserve closer examination.

The purpose of this note is therefore to clarify several common misconceptions about the Swiss-system and to encourage further discussion on possible improvements in tournament design and ranking methods.

Introduction

The Swiss-system is one of the most successful tournament formats ever devised. It allows large competitions to be organised efficiently and has served chess remarkably well for more than a century.

However, its practical success has also created the impression that the system is theoretically well understood and fundamentally sound. This is not entirely the case.

Many widely accepted beliefs about Swiss tournaments are in fact misconceptions. Some concern statistical issues, others arise from structural properties of the pairing mechanism itself. These misunderstandings often influence how tournaments are organised and how final rankings are interpreted.

The following sections discuss several of these misconceptions and explain why certain assumptions about Swiss tournaments deserve closer scrutiny.

Misconception 1: The Swiss system guarantees fair conditions for all players

A common belief is that Swiss tournaments provide essentially equal competitive conditions for all participants. In reality, this was never a design goal of the system.

The Swiss system was created as a practical scheduling mechanism for large tournaments, where a full round-robin is infeasible. Its primary objective is to generate pairings efficiently while approximately matching players with similar scores.

Fairness is only a secondary concern, and only partially enforced. This becomes clear when examining two structural features of the system:

• Colour allocation
• Floating between score groups

These aspects will be discussed in detail in the following sections.

Misconception 2: Players with the same score represent meaningful strength groups

A central idea of the Swiss system is the notion of score groups:
players with the same number of points are treated as roughly equivalent and are typically paired against each other.

This implicitly assumes that the score obtained so far is a reliable proxy for player strength.
In reality, this assumption is fragile for two main reasons.

The arbitrariness of the scoring system

Chess tournaments conventionally assign:

• 1 point for a win
• ½ point for a draw
• 0 points for a loss

This choice is deeply ingrained in practice but not theoretically justified.
Alternative systems (such as 3–1–0) would produce different score distributions and therefore different groupings of players.

Score groups are therefore not intrinsic properties of the competition, but artefacts of a particular scoring convention.

The circularity of pairing and ranking

In the Swiss system, scores are not only used to rank players — they also determine future pairings. This creates a feedback loop:

pairings → results → scores → new pairings

As a consequence, the path through which a player accumulates points matters.
Two players with the same final score may have faced very different opponents and reached that score through very different trajectories.

Artificial groupings

Score groups therefore do not necessarily correspond to meaningful differences in playing strength. They are temporary constructs produced by a small number of games under an arbitrary scoring rule.Nevertheless, the pairing algorithm treats them as if they were meaningful competitive categories.

Misconception 3: Opponent-score tie-breaks measure the strength of opposition

Many tie-break systems used in chess tournaments are based on the scores achieved by a player's opponents. The most common examples are:

• Buchholz
• Median Buchholz
• Buchholz Cut variants
• Sonneborn–Berger

These systems are often described as measuring the strength of the opposition faced by a player. The underlying intuition seems plausible: if a player's opponents score many points in the tournament, then those opponents must have been strong.

However, this interpretation is far more problematic than it appears.

Scores are not independent measurements

In a Swiss tournament, the score obtained by each player is not an independent quantity. It is the result of the pairings generated by the same system. Thus, the scores used to compute Buchholz or similar tie-breaks are themselves the outcome of the pairing process.

This creates a second layer of circularity:

pairings → results → scores → tie-break scores

The tie-break therefore does not measure an external quantity such as the true strength of opponents. It measures the outcome of the same tournament mechanism that produced the pairings. In statistical terms, the system is endogenous.

Sensitivity to unrelated games

Another well-known feature of opponent-score tie-breaks is their sensitivity to results that have nothing to do with the player concerned. If one of your previous opponents wins or loses a game in a later round, your Buchholz score changes — even though your own results remain exactly the same. In extreme cases, the final ranking between two players may depend on the outcome of a game played between two entirely different competitors. This phenomenon is not exceptional; it is an inherent property of opponent-score tie-break systems.

The paradox of the tournament winner

A particularly revealing observation can be made in many Swiss tournaments. It is not uncommon for the winner of the tournament to have a lower Buchholz score than some players who finished behind him in the standings.

This happens for a simple reason. Players who score slightly fewer points may face opponents who later score well in the tournament, while the winner's opponents may include players who lost several games after facing the leader. In other words, the tie-break value is influenced by the future trajectory of opponents, not by their intrinsic playing strength.

The consequence is paradoxical: The player who actually won the tournament may appear to have faced “weaker opposition” according to the tie-break system.

This illustrates that opponent-score tie-breaks are noisy estimators of opponent strength.

From a statistical perspective, systems such as Buchholz can be interpreted as very crude proxies for the strength of opposition. They aggregate quantities that are themselves noisy and path-dependent. As a result, they produce rankings that are sensitive to small perturbations in the tournament results.

This does not mean that such tie-breaks are entirely useless. But it does mean that they should not be interpreted as precise measurements of competitive strength.

A deeper issue

The widespread reliance on opponent-score tie-breaks reveals a deeper structural problem.

The Swiss system produces many ties because it relies on a coarse scoring scale (1, ½, 0) applied over a relatively small number of rounds.

Tie-break systems are therefore required to separate players with identical scores. However, these tie-breaks are built using the same data generated by the tournament itself. They do not introduce genuinely new information. They simply reorganize the existing information through additional layers of calculation.

Misconception 4: Floating rules improve fairness in Swiss tournaments

In the Swiss system, players are normally paired within score groups. However, score groups often contain an odd number of players. When this happens, one player must be paired with an opponent from a different group. This player is called a floater.

The concept of floating is therefore a technical necessity of the Swiss structure. Modern Swiss regulations include numerous restrictions on floating. For example, players are typically prevented from:

• floating in the same direction in consecutive rounds,
• floating again within a short number of rounds,
• or floating under certain score conditions.

These rules are often justified in terms of fairness. But their true role is to stabilize the pairing algorithm and prevent pathological pairing sequences. In effect, they act as constraints that keep the system operational.

A revealing example

Consider a top player who has 3.5 points after round 4 of a tournament. In round 5 he is paired upward against a strong opponent with 4 points, rated 2600, and wins the game. He now has 4.5 points. Suppose that in the next round the group at 5 points contains an odd number of players.

One might expect that the top player — who has just defeated a strong opponent — could again be paired upward into the top group. Yet under the standard floating rules, he may be forbidden to float upward again, simply because he already floated upward in the previous round. This restriction applies regardless of the result of the game he just won.

The algorithm therefore prevents him from playing in the top score group — effectively "protecting" him from facing stronger opponents, even though his previous result gives no reason for such protection.

The Swiss pairing process is therefore not determined solely by the current score distribution. It also depends on historical constraints such as previous floatings and colour allocations. In this sense, the pairing mechanism is path dependent: the sequence of earlier pairings can restrict the set of possible pairings in later rounds.

A structural illusion

Floating rules are therefore built on a fiction already noted in Misconception 2: that score groups represent meaningful strength levels. Treating them as natural competitive categories introduces distortions in the pairing process.

Floating rules do not eliminate these distortions. They merely manage them so that the algorithm can continue to operate.

Misconception 5: An odd number of rounds is preferable in Swiss tournaments

Most large chess tournaments played under the Swiss system use an odd number of rounds. Nine rounds is perhaps the most common format in important tournaments.

This choice is often perceived as natural or even desirable from a sporting perspective. Yet the real reason is much more technical. It is largely related to the difficulty of balancing colours within the Swiss pairing structure.

The colour imbalance problem

In chess, the player with the White pieces enjoys a measurable statistical advantage. Large databases of tournament games consistently show that White scores several percentage points more than Black.

For this reason, tournament formats ideally aim to distribute colours as evenly as possible among players.

In a tournament with an even number of rounds, perfect colour balance is theoretically possible: every player could receive exactly the same number of games with White and Black. For example, in an eight-round tournament every participant could play:

• 4 games with White
• 4 games with Black.

This would ensure identical colour conditions for all players. In practice, however, important Swiss tournaments rarely adopt an even number of rounds. Instead, they typically use an odd number of rounds, such as 9 or 11. The reason lies in the interaction between pairing rules and colour constraints.

The Swiss system simultaneously attempts to:

• pair players within score groups,
• avoid repeated pairings,
• and balance colours.

These objectives can easily come into conflict.

In the later rounds of a tournament — especially the final round — the pairing algorithm often prioritizes pairing players with the same score, even if this creates colour imbalances. FIDE regulations even allow exceptions to colour rules in the last round in order to complete the pairings. In a tournament with an even number of rounds, such exceptions could lead to extreme colour distributions such as: 5 Whites and 3 Blacks or 5 Blacks and 3 Whites. This outcome would clearly introduce a significant structural advantage or disadvantage.

The previous edition of the FIDE Swiss of two decades ago produced the remarkable 6 white/3 blacks distribution in favor of the #1 top GM of an Italian important tournament. This illustrates how extreme colour distortions can arise if the system is not properly constrained.

Thus using an odd number of rounds reduces the probability of such extreme imbalances. But it does not eliminate them.

A structural compromise

The widespread use of nine-round tournaments is therefore not the result of a deep sporting principle. It is a practical compromise designed to limit colour distortions produced by the Swiss pairing mechanism.

Players rarely notice this because colour imbalances are accepted as an unavoidable feature of Swiss tournaments. Yet from a strictly competitive perspective, the situation is problematic.

If White enjoys a measurable advantage, then a player who receives five Whites and four Blacks has enjoyed a better competitive condition than a player who receives the opposite. In other words, the tournament itself introduces unequal conditions among participants.

Traditionally this is justified by the argument that such imbalances average out across many tournaments. But this argument implicitly admits that each individual tournament contains structural biases.

Misconception 6: Tie-break systems create a meaningful ranking

When several players finish a Swiss tournament with the same number of points, tie-break systems are used to determine the final order. FIDE regulations describe these tie-breaks in considerable detail, often specifying long hierarchies of criteria such as:

• Buchholz
• Median Buchholz
• Sonneborn–Berger
• Direct Encounter
• and other variants.

This complexity can give the impression that the final ranking obtained after applying these rules is highly refined and objective. In reality, the situation is far less clear.

The illusion of precision

Tie-break systems attempt to produce a strict ranking from a very limited amount of information. In a typical nine-round Swiss tournament with several hundred players, each participant has played only nine opponents. The final score therefore summarizes the results of a very small sample of games relative to the size of the field.

When several players share the same score, the available information to separate them is extremely limited. Tie-break rules attempt to extract additional distinctions from the same set of results. But this does not create new information. It simply recombines the existing information through additional calculations.

Path dependence

Tie-break values also depend strongly on the pairing path through which a player progressed during the tournament. Two players finishing with the same number of points may have faced opponents with very different trajectories in the event. One player's opponents may subsequently score well, while the other's may perform poorly in later rounds.

The resulting tie-break values therefore reflect not only the strength of opponents but also the random dynamics of the tournament itself. This makes tie-break rankings inherently unstable.

The problem of minimal information

Some tie-break criteria rely on extremely small amounts of information. A typical example is Direct Encounter, which resolves ties based on the result of a single or very few games between the players concerned. Yet from a statistical perspective, a single or very few games provides almost no reliable information about the relative strength of two players.

The result of a single chess game is heavily influenced by:

• opening preparation
• colour allocation
• momentary form
• and random fluctuations in play.

Using such minimal evidence to determine the final ranking of a tournament introduces considerable statistical noise.

A symptom, not a solution

As already noted in Misconception 3, this reliance on tie-breaks reflects a deeper problem: the scoring system produces many ties over a small number of rounds, and no amount of additional criteria can extract information that was never recorded in the first place. The growing catalogue of tie-break rules should therefore not be read as a sign of increasing sophistication, but as an attempt to repair limitations that are built into the Swiss structure itself. They do not eliminate them.

Misconception 7: The Swiss system approximates the ranking of a round-robin tournament

The Swiss system is often informally justified with the following intuition: if strong players keep playing strong opponents, the final ranking should approximate the ranking that would emerge from a full round-robin tournament.

This idea is appealing, but it is largely misleading.

A round-robin provides complete information

In a round-robin tournament every player faces every other player. The final ranking is therefore based on the complete network of results among participants. Although randomness still exists, the structure of the tournament ensures that each competitor is evaluated against the same field of opponents.

The Swiss system operates very differently.

Each player faces only a small subset of opponents, typically fewer than ten in large tournaments with hundreds of participants. The information available about relative strength is therefore extremely limited.

Path dependence

In Swiss tournaments the sequence of pairings strongly influences the final outcome. A player who happens to face strong opponents early may lose points and drop into lower score groups, where the level of opposition is weaker.

Conversely, another player may accumulate points against weaker opponents and remain in higher score groups throughout the tournament. Once these trajectories diverge, they tend to reinforce themselves. This phenomenon is sometimes described as path dependence.

Two players of similar strength can therefore experience very different tournament paths depending on early results and pairings. The final ranking may reflect these trajectories as much as the actual strength of the players.

The limited convergence of Swiss tournaments

The idea that Swiss tournaments naturally converge toward the true ranking of players would require a large number of rounds. In practice, however, Swiss tournaments are typically very short relative to the number of participants.

For example, in a nine-round Swiss tournament with 200 players, each participant plays fewer than 5% of the possible opponents. Under such conditions, the ranking cannot reliably approximate the full ordering that would emerge from a round-robin competition.

Instead, it represents a partial ordering produced by a particular sequence of pairings.

The role of tie-breaks

Tie-break systems attempt to compensate for this lack of information. By incorporating the scores of opponents or other derived quantities, they try to infer additional distinctions among players. But as discussed earlier, these tie-breaks are themselves based on the same limited and path-dependent data. They do not fundamentally resolve the problem.

A different perspective

From a modern statistical viewpoint, the results of a tournament form a network of games. Each game provides partial information about the relative strengths of two players. The goal of ranking should be to extract the most coherent estimate of player strength from this entire network.

Swiss tournaments do produce such a network.

But the traditional ranking method — based on point totals and tie-break rules — does not use the available information in a fully coherent way.

Misconception 8: Top-vs-Bottom pairings create balanced encounters and mix players well

A common belief about the Swiss system is that the top-vs-bottom rule inside score groups produces balanced encounters and mixes players effectively. In reality, its structural effect is quite different.

Within each score group, players are ordered according to the initial rating. The group is then split into two halves and players from the top half are paired with players from the bottom half.

This rule introduces a systematic asymmetry that players are aware of: higher-ranked players within the group tend to face slightly lower-ranked opponents. As a result, the pairing mechanism tends to preserve the initial ordering rather than neutralize it.

This effect becomes particularly visible when two players obtain identical results throughout the tournament. Suppose players #1 and #2 score exactly the same number of points and draw their direct encounter. In such a situation, player #1 will almost inevitably remain ahead of player #2 in the final ranking under most standard tie-break systems (such as Buchholz or Sonneborn-Berger).

The reason is structural. Because of the top-vs-bottom rule, player #1 tends to face opponents who are slightly higher ranked than those of player #2. Consequently, the tie-break calculations systematically favour the initially higher-ranked player.

In this sense, the Swiss system never completely “forgets” the initial ranking. Even when players share the same score, the pairing procedure continues to reflect the original ordering.

Thus the widely held belief that the Swiss system fully resets competition within each score group is misleading. In practice, the initial ranking acts as a persistent hidden ordering mechanism throughout the tournament.

Misconception 9: The Swiss system naturally produces a decisive final round

A common belief is that Swiss tournaments tend to produce a dramatic final round in which the winner of the event is decided directly over the board.

The intuition seems appealing: if the strongest players gradually rise to the top of the standings, they should eventually meet in the last round, creating a natural “final”.

In practice, however, the situation is often quite different.

Because Swiss tournaments rank players primarily by points, even a small lead in the standings can strongly influence players’ incentives. A player who enters the final round with a half-point advantage often has little reason to take risks. In many situations a draw is sufficient to secure first place, especially when favourable tie-breaks are expected.

As a result, the supposedly decisive final-round encounter may be played with a fundamentally conservative strategy.

This dynamic differs markedly from true finals in other sports. In a tennis final, for example, both players must compete fully for victory; a draw is impossible and neither player benefits from avoiding risk.

In Swiss tournaments, by contrast, the asymmetry created by the standings can reduce the competitive intensity of the final game. The player in the lead may rationally choose safety over ambition.

For this reason, the idea that the Swiss system naturally produces a dramatic and decisive final round is not always justified.

Misconception 10: The Swiss system is theoretically well understood

Given the level of detail contained in modern tournament regulations, it is easy to assume that the Swiss system rests on a well-established theoretical foundation.

The pairing rules are carefully specified, the floating procedures are precisely defined, and the hierarchy of tie-break systems is described in extensive detail.

This apparent precision can create the impression that the system is the result of a coherent mathematical theory. In reality, the Swiss system developed in a very different way.

A practical invention

The Swiss system was introduced in the late nineteenth century as a practical solution to a logistical problem: how to organize large tournaments without requiring a full round-robin schedule.

Its goal was operational rather than theoretical. Tournament organizers needed a method that could:

• produce pairings quickly,
• approximately match players with similar scores,
• and allow a tournament to finish within a limited number of rounds.

Within those constraints, the Swiss system proved remarkably effective. But it was not derived from a theory of ranking or from a formal model of competition.

Over time, new practical issues emerged. Organizers and arbiters observed problems such as:

• colour imbalances,
• unstable rankings,
• and pathological pairing sequences.

To address these issues, additional rules were gradually introduced. Floating constraints, colour preferences, pairing priorities, and increasingly elaborate tie-break systems were added one after another. The result is the modern Swiss system: a large collection of procedural rules designed to handle the many corner cases that arise in practice.

This process resembles engineering refinement rather than theoretical design.

The software era

As the system accumulated rules and exceptions, its implementation became increasingly complex. Modern Swiss pairing procedures are now sufficiently intricate that they are almost always executed using specialized software. In straightforward situations the procedure can be followed manually, but once moderately complex cases arise only a small number of arbiters are able to reproduce the full pairing process without software assistance. This is not a criticism of arbiters, but a reflection of how the Swiss system has evolved over time.

This illustrates an interesting paradox. A system originally designed for practical simplicity has gradually evolved into a highly technical algorithm whose behaviour is understood mainly through its software implementations.

An empirical tradition

None of this diminishes the historical success of the Swiss system. For more than a century it has allowed chess tournaments to function on a large scale. However, it is important to recognize the nature of the system.

The Swiss system is not the outcome of a unified theoretical framework for ranking competitors under incomplete information. It is the result of a long empirical tradition in tournament organization.

Recognizing this distinction is the first step toward asking a natural question:

If we were to design a tournament system today, using the mathematical and statistical tools now available, would we design it in the same way?

What Should Replace the Current Swiss System?

Criticizing the limitations of the current Swiss system does not imply that it should simply be abandoned.

For more than a century, the Swiss format has allowed the chess world to organize large tournaments efficiently and at relatively low cost. Its historical importance should not be underestimated.

The question is therefore not whether Swiss tournaments should disappear, but whether their underlying principles can be rethought and improved.

Preserving what works and removing unnecessary constraints

While the Swiss system has important limitations, it also contains valuable ideas that should be preserved. At the same time, some of its constraints originate from historical circumstances that are no longer relevant today. The following principles aim to retain the strengths of the traditional system while removing unnecessary restrictions that limit the design of improved tournament formats.

1. Pairings must be deterministic.

Given the initial conditions of a tournament — the list of players, ratings, colours, and previous results — the pairing algorithm should always produce the same pairings.

This requirement is crucial for transparency and sporting fairness. Random pairing procedures might sometimes have attractive statistical properties, but they would introduce unacceptable risks in competitive environments.

Determinism protects tournaments from:

• manipulation,
• favoritism,
• and suspicions that certain players might receive artificially favourable pairings.

For this reason, any future pairing system must remain fully reproducible and verifiable.

2. Separating pairing from ranking

The most important conceptual step is to recognize that pairing and ranking are two different problems.

• The pairing system determines which games are played.
• The ranking system determines how the results of those games are interpreted.

In the traditional Swiss system these functions are tightly intertwined. But there is no fundamental reason why they must be.

Once these roles are separated, both tasks can be addressed more coherently. Pairing algorithms can focus on:

• fairness of conditions,
• colour balance,
• structural stability of the tournament.

Ranking methods, on the other hand, can use the entire network of games to estimate the relative strength of players. Modern statistical models already provide powerful tools for this purpose.

3. Pairing decisions should depend only on the current tournament state, not on historical artefacts of the pairing algorithm.

This principle limits the path dependence of the pairing mechanism. If pairing decisions depend on historical artefacts such as floating history, the tournament state is no longer determined solely by current results but also by past algorithmic constraints. This can distort the competitive structure of later rounds.

4. Manual executability should not be considered a fundamental requirement.

Traditional Swiss systems were designed so that arbiters could generate pairings without computer assistance. In modern tournaments this constraint has largely lost its practical relevance. While manual executability may remain desirable, it should not limit the design of improved pairing systems if relaxing this requirement allows fairer and more coherent competitions.

A more coherent use of information

In a tournament, every game contributes information about the relative strength of the players involved. Traditional Swiss rankings use this information only partially, through point totals and tie-break systems. Statistical ranking models can instead use the full structure of the game network, producing strength estimates that are more stable and less sensitive to arbitrary tie-break rules.

In such a framework, ties become much rarer and the need for complex tie-break hierarchies largely disappears.

Toward fairer tournaments

Separating pairing and ranking also makes it easier to enforce conditions that are difficult to guarantee within the traditional Swiss structure. For example:

• strict balance of colours,
• symmetric competitive conditions for all players,
• and more robust final rankings.

Such improvements would not require abandoning Swiss-style tournaments. They would simply require rethinking their internal architecture.

An open project

The purpose of this project is not to impose a single definitive solution. Instead, it is to open a discussion that connects the chess community with the modern scientific literature on ranking systems and competition design.

The Swiss system was a brilliant practical invention for its time. But more than a century later, chess has the opportunity to combine its rich tournament tradition with the mathematical tools developed in the intervening decades.

The future of chess tournaments should not be determined only by tradition. It should also reflect the knowledge now available.

Before concluding, it is important to acknowledge the work of the many arbiters, programmers, and researchers who have devoted enormous effort to developing and refining Swiss pairing systems over the years. Modern chess tournaments rely heavily on the algorithms and software created by these contributors, often developed freely and shared with the community, and the global growth of competitive chess owes a great deal to that effort. The remarkable success of the Swiss system is itself evidence of how effective this work has been in practice.

The ideas presented here should therefore be understood as an invitation rather than a criticism: an invitation to continue the same tradition of open collaboration that has supported tournament chess for more than a century, by combining the practical experience of arbiters and programmers with the analytical tools now available in statistics and tournament theory. The goal is not to abandon the achievements of the past, but to build upon them.

• revision 5: 2026-06-16
• revision 4: 2026-04-08
• revision 3: 2026-03-14
• revision 2: 2026-03-11
• revision 1: 2026-03-07